Aggregator

At the end of 2017, we named Evidence Action's No Lean Season one of GiveWell's nine top charities. Now, GiveWell and Evidence Action agree that No Lean Season should not be a GiveWell top charity this year, and Evidence Action is not seeking additional funding to support No Lean Season's work at this time.

This post will discuss this decision in detail. In brief, we updated our assessment of No Lean Season, a program that provides loans to support seasonal migration, based on preliminary results Evidence Action began discussing with us in July from a study of the 2017 implementation of the program (hereinafter referred to as “2017 RCT”). These results suggested the program, as implemented in 2017, did not successfully induce migration. Taking this new information into account alongside previous studies of the program, we and Evidence Action do not believe No Lean Season meets our top charity criteria at this time.

Evidence Action's post on this decision is here.

GiveWell's mission is to identify and recommend charities that can most effectively use additional donations. While it may be disappointing for a top charity to be removed from our list of recommendations, we believe that adding and removing top charities from our list is an important part of our process. If our top charities list never changed, we would guess we were (a) acting too conservatively (i.e. not being open enough to adding new top charities), or (b) not being critical enough of groups once they've been added to our list (i.e. not being open enough to removing existing top charities).

We believe this decision speaks positively of Evidence Action and demonstrates our mutual commitment to updating our views based on new evidence. GiveWell has interacted with hundreds of organizations in our history, and very few have subjected their programs to a rigorous study in the way that Evidence Action did last year and, at smaller scale, in 2014. We're excited to work with a group like Evidence Action that is committed to rigorous study and openness about results.

Summary

In this post, we will discuss:

• The history of GiveWell and No Lean Season. (More)
• How the 2017 RCT updated our views of No Lean Season. (More)
  • What did the 2017 RCT find? (More)
  • How did we interpret the RCT results? (More)
  • What does the future of No Lean Season look like? (More)
• Conclusion

Read More

The post Update on No Lean Season’s top charity status appeared first on The GiveWell Blog.

At the end of 2017, we named Evidence Action’s No Lean Season one of GiveWell’s nine top charities. Now, GiveWell and Evidence Action agree that No Lean Season should not be a GiveWell top charity this year, and Evidence Action is not seeking additional funding to support its work at this time.

This post will discuss this decision in detail. In brief, we updated our assessment of No Lean Season, a program that provides loans to support seasonal migration, based on preliminary results Evidence Action began discussing with us in July from a study of the 2017 implementation of the program (hereinafter referred to as “2017 RCT”). These results suggested the program, as implemented in 2017, did not successfully induce migration. Taking this new information into account alongside previous studies of the program, we and Evidence Action do not believe No Lean Season meets our top charity criteria at this time.

Evidence Action’s post on this decision is here.

GiveWell’s mission is to identify and recommend charities that can most effectively use additional donations. While it may be disappointing for a top charity to be removed from our list of recommendations, we believe that adding and removing top charities from our list is an important part of our process. If our top charities list never changed, we would guess we were (a) acting too conservatively (i.e. not being open enough to adding new top charities), or (b) not being critical enough of groups once they’ve been added to our list (i.e. not being open enough to removing existing top charities).

We believe this decision speaks positively of Evidence Action and demonstrates our mutual commitment to updating our views based on new evidence. GiveWell has interacted with hundreds of organizations in our history, and very few have subjected their programs to a rigorous study in the way that Evidence Action did last year and, at smaller scale, in 2014. We’re excited to work with a group like Evidence Action that is committed to rigorous study and openness about results.

Summary

In this post, we will discuss:

• The history of GiveWell and No Lean Season. (More)
• How the 2017 RCT updated our views of No Lean Season. (More)
  • What did the 2017 RCT find? (More)
  • How did we interpret the RCT results? (More)
  • What does the future of No Lean Season look like? (More)
• Conclusion

GiveWell and No Lean Season

No Lean Season provides support for low-income agricultural workers in rural Bangladesh during the time of seasonal income and food insecurity (“lean season”). The program provides small, interest-free loans to support workers’ temporary migration to seek employment. No Lean Season is implemented by RDRS Bangladesh; Evidence Action provides strategic direction, conducts program monitoring, and provides technical assistance, among other functions. Evidence Action developed No Lean Season as part of its Beta portfolio, which is focused on prototyping and scaling cost-effective programs.

GiveWell began engaging with No Lean Season as a potential top charity in 2013, when we began to explore making an Incubation Grant to support its scale-up. We saw No Lean Season as a promising program that lacked the track record to be considered for a top charity recommendation at that time. We describe our initial interest in the program in a February 2017 blog post:

We approached Evidence Action in late 2013 to express our interest in supporting the creation of new GiveWell top charities.

In March 2014, Good Ventures made a $250,000 grant to Evidence Action to support the investigation and scale-up of promising programs.

Since then, Good Ventures has made three additional grants totaling approximately $2.7 million to support the program’s scale-up.

No Lean Season continued to test and scale their program with this and other support. We decided to recommend No Lean Season as a top charity in late 2017. We based our recommendation on three randomized controlled trials (RCTs) of the program. (We generally consider RCTs to be one of the strongest types of evidence available; you can read more about why we rely on RCTs here.)

Two of the RCTs (conducted in 2008 and 2014) indicated increased migration, income, and consumption for program participants. In the third RCT, which was conducted in 2013 and has not been published, the program is considered to have failed to induce migration, potentially due to political violence that year. We discuss the RCT evidence in greater depth in our intervention report on conditional subsidies for seasonal labor migration in northern Bangladesh.

Weighing the evidence, the cost of the program, and the potential impacts, we decided No Lean Season met our criteria to be named a top charity in November 2017. We summarized our reasoning in our blog post announcing our 2017 list of top charities, and noted the risks of this recommendation:

Several randomized controlled trials (RCTs) of subsidies to increase migration provide moderately strong evidence that such an intervention increases household income and consumption during the lean season. An additional RCT is ongoing. We estimate that No Lean Season is roughly five times as cost-effective as cash transfers (see our cost-effectiveness analysis).

Evidence Action has shared some details of its plans for monitoring No Lean Season in the future, but, as many of these plans have not been fully implemented, we have seen limited results. Therefore, there is some uncertainty as to whether No Lean Season will produce the data required to give us confidence that loans are appropriately targeted and reach their intended recipients in full; that recipients are not pressured into accepting loans; and that participants successfully migrate, find work, and are not exposed to major physical and other risks while migrating.

As indicated above, No Lean Season conducted an additional RCT to evaluate its program during the 2017 lean season (approximately September to December), the preliminary results of which indicate the program failed to induce migration. With the evidence from the 2017 RCT, the case for the program’s impact and cost-effectiveness looks weaker.

Our updated perspective on No Lean Season

The 2017 RCT was a key factor in the decision to remove No Lean Season from our top charities list. Below, we discuss:

• What the 2017 RCT found.
• Our interpretation of the RCT results.
• No Lean Season’s future.

What did the 2017 RCT find?

The 2017 RCT was a collaboration between Evidence Action, Innovations for Poverty Action, and researchers from Yale University, the London School of Economics, and the University of California, Davis. In a preliminary analysis shared with GiveWell in September 2018, the researchers did not find evidence for a negative or positive impact on migration, and found no statistically significant impact on income and consumption.[1]

However, the implementation of the program during the 2017[2] lean season and the evaluation of it differed from previous iterations. No Lean Season operated at a larger scale in the fall of 2017 than it had previously, offering loans to 158,155 households, compared with 16,268 households in 2016. Relative to earlier versions of the program, the program in 2017 involved (a) higher-intensity delivery of the intervention (offering loans to most eligible individuals) and (b) broader eligibility requirements (the eligibility rate in 2017 was 77 percent, compared with 49 percent in 2016).[3]

At this point, neither GiveWell, nor No Lean Season, nor the researchers feel we have a conclusive understanding of why the program failed to induce migration. However, No Lean Season and the researchers are exploring various hypotheses about what may explain the failure to induce migration, and they note that some suggestive evidence supports some hypotheses more than others. The researchers have posited several possibilities:

1. The way the program was targeted in 2017 was suboptimal. The Migration Organizers, who survey households for eligibility and offer and disburse loans (more detail here under “Migration Organizers”), may have focused their efforts on the individuals that were seen as most likely to migrate, rather than those who needed a loan to afford migration. The use of loan targets during implementation may have inadvertently incentivized this behavior.[4] If, for example, loan officers mostly made loans to people who would have migrated regardless of receiving a loan, this could have led to the lack of impact on migration found in the study.
2. The 2017 lean season was particularly bad for the program. The researchers note that severe flooding and associated implementation delays in some regions may have caused problems in 2017. The researchers plan to look more closely at the regions that experienced flooding, though they note that they don’t have the data necessary to make experimental comparisons.[5] In addition, a 2013 trial may have failed due to issues that were specific to the year of that trial, such as increased labor strikes.
3. There exists another (currently unknown) reason why this program won’t work at scale. Conditions in Bangladesh may have changed, negative spillovers (harmful impacts for individuals who did not receive loans) may cancel out gains, or pilot villages may have been strategically picked in earlier trials.[6]

The researchers are considering all of these possibilities. After considering various possible theories as well as some non-experimental data (including administrative data and data from a special-purpose survey of Migration Organizers who worked on the program in 2017), they feel that the ‘mistargeting’ theory is the most likely explanation and the explanation most consistent with the analysis.[7]

In scenario (1), No Lean Season may be able to identify and fix the problem. In scenario (2), GiveWell will need to update our estimate of the impact of the program to take into account the fact that periodic program failures due to external factors are more likely than we previously thought. In scenario (3), the program is unlikely to be effective in the future.

How did we interpret the RCT results?

We don’t know the extent to which each of the above explanations contributed to the study not finding an effect on migration.

We used the results of the 2017 RCT to update our cost-effectiveness estimate for the program. Cost-effectiveness estimates form arguably the most important single input into our decisions about whether or not to recommend charities (more on how GiveWell uses cost-effectiveness analyses here). When we calculate a program’s cost-effectiveness, we take many different factors into account, such as the administrative and program costs and the expected impact. We also make a number of educated guesses, such as the likelihood that a program’s impact in a new country will be similar to that in a country where it has previously worked. Below, we describe the mechanism by which the 2017 RCT result was incorporated into our model and how it changed our conclusion.

Prior to this year, we formed our view of No Lean Season based on the three small-scale RCTs mentioned above (conducted in 2008, 2013, and 2014). Each of these RCTs looked at a slightly different version of the program. We believed that the ‘high-intensity’ arm of the 2014 RCT was the version most likely to resemble the program at scale. We thus used the migration rate measured in this arm of the RCT as our starting point for calculating the program’s impact.

The high-intensity arm of the 2014 RCT also had the highest measured migration rate of the three RCTs we assessed, and so we wanted to give some consideration to the less-positive results found in the other two assessments. We applied a small, downward adjustment to the rate of induced migration observed in the 2014 high-intensity arm in our cost-effectiveness model; this was an educated guess, based on the information we had. Our best guess was that the program would lead, in expectation, to 80% of the induced migration seen in the 2014 high-intensity arm.[8]

Now, the preliminary 2017 RCT results show no significant impact on migration rates or incomes. Because this trial was large and very recent, we updated our expectations of the impact of the program substantially, and in a negative direction. Our best guess now is that the program will lead, in expectation, to 40% of the induced migration seen in the 2014 high-intensity arm. Holding other inputs constant, this adjustment reduces our estimate of No Lean Season’s cost-effectiveness by a factor of two.

This reduced cost-effectiveness, along with our updated qualitative picture of No Lean Season’s evidence of effectiveness, led to the decision to remove No Lean Season from our top charities list.

What does the future of No Lean Season look like?

Although they are not raising more funding at this time, No Lean Season has over two years’ worth of remaining funding. We understand that the organization has made changes to the program design in 2018 based on emerging interpretations of the 2017 results, and has collected additional data to evaluate some of the hypotheses which may explain those results (including, for example, a survey of Migration Organizers who worked on the 2017 program). They plan to subject the 2018 implementation round to an additional ‘RCT-at-scale,’ with a particular focus on reassessing the program’s effects on migration, income and consumption, as well as potential effects at migration destinations. They will continue to explore what may have caused the issue in the 2017 program at scale, and to see whether they can find a solution. If they do that, we’ll want to reassess the evidence and the costs to determine whether No Lean Season meets our bar for top charity status. Evidence Action believes we should have the necessary information to reassess starting in mid-2019, based on the results of the RCT conducted during the 2018 lean season and other analyses they perform.

Conclusion

This is the second time since 2011 that we have removed a top charity from our list (prior to 2011, our top charities list was fairly different from today; we made a big-picture shift in our priorities that year that led us to our more recent lists). The previous removal occurred in 2013, when we took the Against Malaria Foundation (AMF) off of our list because we didn’t believe it could absorb additional funding effectively in the near term. AMF was reinstated as a top charity in 2014.

The decision to remove a top charity is never easy. But continuously evaluating GiveWell’s recommended charities is an important part of our work, and we take it seriously. It’s easy to talk about a commitment to evidence when the results are positive. It’s hard to maintain that commitment when the results are not. We’re excited to work with a group like Evidence Action that is committed to rigorous program evaluation and open discussion of the results of those evaluations. Its openness about these results has increased our confidence in Evidence Action as an organization. We look forward to seeing the results from the 2018 RCT in 2019.

Notes

[1] “At this early stage in analysis, we find no evidence that the program had an impact (positive or negative) on migration, caloric intake, food expenditure, or income.” Evidence Action, unpublished summary document, Page 1.

[2] The 2017 RCT studied a period from the fall of 2017 through early 2018.

[3] “This study has two main goals:

1. “A replication of previous findings showing positive impact of incentivized migration on seasonal migration, caloric intake, food and non-food expenditure, income, and food security. Our aim is to estimate impact of a scaled version of the No Lean Season program: intensifying program implementation within branches and expanding the provision of loans to all eligible households.”

Unpublished summary document, Page 1.

[4] “The second set of explanations focus on unintentional implementation changes caused by the change ineligibility, the vastly expanded scope of the program, or other factors. In the most recent round, it is possible that Migration Organizers (MOs) focused their efforts on those households who were most likely to migrate even without a loan to the exclusion of the target population households who need a loan to afford migration. Such behavior may have even been encouraged by the use of targets set by the NGO to manage implementation at such a large scale. We have implemented a qualitative survey to understand the incentives and actions of MOs last year, and are revising our instructions to avoid any possibility of this issue this year.” Evidence Action, unpublished summary document (with minor revision from Evidence Action), Page 11.

[5] “Most notably, the program was affected by severe flooding in many regions, and implementation was subsequently delayed as well. We are still evaluating whether these regions are the ones with the most diminished effects, although we lack the data in control areas to conduct an experimental comparison.” Evidence Action, unpublished summary document, Page 11-12.

[6] “It is possible that what we observe this year may be the true effect of the No Lean Season program when implemented at scale. This may be because conditions in rural Bangladesh have changed since the initial years of success, spillovers at scale cancel out any gains observed in small-scale pilots, or pilot villages were selected because they were most likely to be receptive to the program.” Evidence Action, unpublished summary document, Page 11.

[7] Evidence Action, “Interpretation of 2017 Results” deck and narrative (unpublished)

[8] “This adjustment is used to account for external validity concerns not accounted for elsewhere in the CEA.

“The default adjustment value of 80% is our best guess about the appropriate value, but it is not based on a formal calculation.

“The program at scale takes place in the same region with the same implementers (RDRS and Evidence Action) as the source of our key evidence for the intervention (the 2014 RCT). The program at scale differs in some aspects of implementation, particularly the inclusiveness of the eligibility criteria and the proportion of eligible households offered an incentive. In the 2014 RCT, the subsidy was a cash transfer rather than an interest-free loan, however the 2008 RCT found a similar effect regardless of whether the subsidy was a cash transfer or an interest-free loan.

“There is some evidence (from a 2013 RCT) suggesting that the program may be ineffective when the perceived risk of migrating increases for reasons such as labor strikes and violence. The researchers estimated that these are 1-in-10 year events.

“Additional discussion related to this parameter can be found at https://www.givewell.org/charities/no-lean-season#programdifferentfromRCTs.” 2018 GiveWell Cost-Effectiveness Model — Version 10, “Migration subsidies” tab, note on cell A19.

The post Update on No Lean Season’s top charity status appeared first on The GiveWell Blog.

At the end of 2017, we named Evidence Action’s No Lean Season one of GiveWell’s nine top charities. Now, GiveWell and Evidence Action agree that No Lean Season should not be a GiveWell top charity this year, and Evidence Action is not seeking additional funding to support its work at this time.

This post will discuss this decision in detail. In brief, we updated our assessment of No Lean Season, a program that provides loans to support seasonal migration, based on preliminary results Evidence Action began discussing with us in July from a study of the 2017 implementation of the program (hereinafter referred to as “2017 RCT”). These results suggested the program, as implemented in 2017, did not successfully induce migration. Taking this new information into account alongside previous studies of the program, we and Evidence Action do not believe No Lean Season meets our top charity criteria at this time.

Evidence Action’s post on this decision is here.

GiveWell’s mission is to identify and recommend charities that can most effectively use additional donations. While it may be disappointing for a top charity to be removed from our list of recommendations, we believe that adding and removing top charities from our list is an important part of our process. If our top charities list never changed, we would guess we were (a) acting too conservatively (i.e. not being open enough to adding new top charities), or (b) not being critical enough of groups once they’ve been added to our list (i.e. not being open enough to removing existing top charities).

We believe this decision speaks positively of Evidence Action and demonstrates our mutual commitment to updating our views based on new evidence. GiveWell has interacted with hundreds of organizations in our history, and very few have subjected their programs to a rigorous study in the way that Evidence Action did last year and, at smaller scale, in 2014. We’re excited to work with a group like Evidence Action that is committed to rigorous study and openness about results.

Summary

In this post, we will discuss:

• The history of GiveWell and No Lean Season. (More)
• How the 2017 RCT updated our views of No Lean Season. (More)
  • What did the 2017 RCT find? (More)
  • How did we interpret the RCT results? (More)
  • What does the future of No Lean Season look like? (More)
• Conclusion

GiveWell and No Lean Season

No Lean Season provides support for low-income agricultural workers in rural Bangladesh during the time of seasonal income and food insecurity (“lean season”). The program provides small, interest-free loans to support workers’ temporary migration to seek employment. No Lean Season is implemented by RDRS Bangladesh; Evidence Action provides strategic direction, conducts program monitoring, and provides technical assistance, among other functions. Evidence Action developed No Lean Season as part of its Beta portfolio, which is focused on prototyping and scaling cost-effective programs.

GiveWell began engaging with No Lean Season as a potential top charity in 2013, when we began to explore making an Incubation Grant to support its scale-up. We saw No Lean Season as a promising program that lacked the track record to be considered for a top charity recommendation at that time. We describe our initial interest in the program in a February 2017 blog post:

We approached Evidence Action in late 2013 to express our interest in supporting the creation of new GiveWell top charities.

In March 2014, Good Ventures made a $250,000 grant to Evidence Action to support the investigation and scale-up of promising programs.

Since then, Good Ventures has made three additional grants totaling approximately $2.7 million to support the program’s scale-up.

No Lean Season continued to test and scale their program with this and other support. We decided to recommend No Lean Season as a top charity in late 2017. We based our recommendation on three randomized controlled trials (RCTs) of the program. (We generally consider RCTs to be one of the strongest types of evidence available; you can read more about why we rely on RCTs here.)

Two of the RCTs (conducted in 2008 and 2014) indicated increased migration, income, and consumption for program participants. In the third RCT, which was conducted in 2013 and has not been published, the program is considered to have failed to induce migration, potentially due to political violence that year. We discuss the RCT evidence in greater depth in our intervention report on conditional subsidies for seasonal labor migration in northern Bangladesh.

Weighing the evidence, the cost of the program, and the potential impacts, we decided No Lean Season met our criteria to be named a top charity in November 2017. We summarized our reasoning in our blog post announcing our 2017 list of top charities, and noted the risks of this recommendation:

Several randomized controlled trials (RCTs) of subsidies to increase migration provide moderately strong evidence that such an intervention increases household income and consumption during the lean season. An additional RCT is ongoing. We estimate that No Lean Season is roughly five times as cost-effective as cash transfers (see our cost-effectiveness analysis).

Evidence Action has shared some details of its plans for monitoring No Lean Season in the future, but, as many of these plans have not been fully implemented, we have seen limited results. Therefore, there is some uncertainty as to whether No Lean Season will produce the data required to give us confidence that loans are appropriately targeted and reach their intended recipients in full; that recipients are not pressured into accepting loans; and that participants successfully migrate, find work, and are not exposed to major physical and other risks while migrating.

As indicated above, No Lean Season conducted an additional RCT to evaluate its program during the 2017 lean season (approximately September to December), the preliminary results of which indicate the program failed to induce migration. With the evidence from the 2017 RCT, the case for the program’s impact and cost-effectiveness looks weaker.

Our updated perspective on No Lean Season

The 2017 RCT was a key factor in the decision to remove No Lean Season from our top charities list. Below, we discuss:

• What the 2017 RCT found.
• Our interpretation of the RCT results.
• No Lean Season’s future.

What did the 2017 RCT find?

The 2017 RCT was a collaboration between Evidence Action, Innovations for Poverty Action, and researchers from Yale University, the London School of Economics, and the University of California, Davis. In a preliminary analysis shared with GiveWell in September 2018, the researchers did not find evidence for a negative or positive impact on migration, and found no statistically significant impact on income and consumption.[1]

However, the implementation of the program during the 2017[2] lean season and the evaluation of it differed from previous iterations. No Lean Season operated at a larger scale in the fall of 2017 than it had previously, offering loans to 158,155 households, compared with 16,268 households in 2016. Relative to earlier versions of the program, the program in 2017 involved (a) higher-intensity delivery of the intervention (offering loans to most eligible individuals) and (b) broader eligibility requirements (the eligibility rate in 2017 was 77 percent, compared with 49 percent in 2016).[3]

At this point, neither GiveWell, nor No Lean Season, nor the researchers feel we have a conclusive understanding of why the program failed to induce migration. However, No Lean Season and the researchers are exploring various hypotheses about what may explain the failure to induce migration, and they note that some suggestive evidence supports some hypotheses more than others. The researchers have posited several possibilities:

1. The way the program was targeted in 2017 was suboptimal. The Migration Organizers, who survey households for eligibility and offer and disburse loans (more detail here under “Migration Organizers”), may have focused their efforts on the individuals that were seen as most likely to migrate, rather than those who needed a loan to afford migration. The use of loan targets during implementation may have inadvertently incentivized this behavior.[4] If, for example, loan officers mostly made loans to people who would have migrated regardless of receiving a loan, this could have led to the lack of impact on migration found in the study.
2. The 2017 lean season was particularly bad for the program. The researchers note that severe flooding and associated implementation delays in some regions may have caused problems in 2017. The researchers plan to look more closely at the regions that experienced flooding, though they note that they don’t have the data necessary to make experimental comparisons.[5] In addition, a 2013 trial may have failed due to issues that were specific to the year of that trial, such as increased labor strikes.
3. There exists another (currently unknown) reason why this program won’t work at scale. Conditions in Bangladesh may have changed, negative spillovers (harmful impacts for individuals who did not receive loans) may cancel out gains, or pilot villages may have been strategically picked in earlier trials.[6]

The researchers are considering all of these possibilities. After considering various possible theories as well as some non-experimental data (including administrative data and data from a special-purpose survey of Migration Organizers who worked on the program in 2017), they feel that the ‘mistargeting’ theory is the most likely explanation and the explanation most consistent with the analysis.[7]

In scenario (1), No Lean Season may be able to identify and fix the problem. In scenario (2), GiveWell will need to update our estimate of the impact of the program to take into account the fact that periodic program failures due to external factors are more likely than we previously thought. In scenario (3), the program is unlikely to be effective in the future.

How did we interpret the RCT results?

We don’t know the extent to which each of the above explanations contributed to the study not finding an effect on migration.

We used the results of the 2017 RCT to update our cost-effectiveness estimate for the program. Cost-effectiveness estimates form arguably the most important single input into our decisions about whether or not to recommend charities (more on how GiveWell uses cost-effectiveness analyses here). When we calculate a program’s cost-effectiveness, we take many different factors into account, such as the administrative and program costs and the expected impact. We also make a number of educated guesses, such as the likelihood that a program’s impact in a new country will be similar to that in a country where it has previously worked. Below, we describe the mechanism by which the 2017 RCT result was incorporated into our model and how it changed our conclusion.

Prior to this year, we formed our view of No Lean Season based on the three small-scale RCTs mentioned above (conducted in 2008, 2013, and 2014). Each of these RCTs looked at a slightly different version of the program. We believed that the ‘high-intensity’ arm of the 2014 RCT was the version most likely to resemble the program at scale. We thus used the migration rate measured in this arm of the RCT as our starting point for calculating the program’s impact.

The high-intensity arm of the 2014 RCT also had the highest measured migration rate of the three RCTs we assessed, and so we wanted to give some consideration to the less-positive results found in the other two assessments. We applied a small, downward adjustment to the rate of induced migration observed in the 2014 high-intensity arm in our cost-effectiveness model; this was an educated guess, based on the information we had. Our best guess was that the program would lead, in expectation, to 80% of the induced migration seen in the 2014 high-intensity arm.[8]

Now, the preliminary 2017 RCT results show no significant impact on migration rates or incomes. Because this trial was large and very recent, we updated our expectations of the impact of the program substantially, and in a negative direction. Our best guess now is that the program will lead, in expectation, to 40% of the induced migration seen in the 2014 high-intensity arm. Holding other inputs constant, this adjustment reduces our estimate of No Lean Season’s cost-effectiveness by a factor of two.

This reduced cost-effectiveness, along with our updated qualitative picture of No Lean Season’s evidence of effectiveness, led to the decision to remove No Lean Season from our top charities list.

What does the future of No Lean Season look like?

Although they are not raising more funding at this time, No Lean Season has over two years’ worth of remaining funding. We understand that the organization has made changes to the program design in 2018 based on emerging interpretations of the 2017 results, and has collected additional data to evaluate some of the hypotheses which may explain those results (including, for example, a survey of Migration Organizers who worked on the 2017 program). They plan to subject the 2018 implementation round to an additional ‘RCT-at-scale,’ with a particular focus on reassessing the program’s effects on migration, income and consumption, as well as potential effects at migration destinations. They will continue to explore what may have caused the issue in the 2017 program at scale, and to see whether they can find a solution. If they do that, we’ll want to reassess the evidence and the costs to determine whether No Lean Season meets our bar for top charity status. Evidence Action believes we should have the necessary information to reassess starting in mid-2019, based on the results of the RCT conducted during the 2018 lean season and other analyses they perform.

Conclusion

This is the second time since 2011 that we have removed a top charity from our list (prior to 2011, our top charities list was fairly different from today; we made a big-picture shift in our priorities that year that led us to our more recent lists). The previous removal occurred in 2013, when we took the Against Malaria Foundation (AMF) off of our list because we didn’t believe it could absorb additional funding effectively in the near term. AMF was reinstated as a top charity in 2014.

The decision to remove a top charity is never easy. But continuously evaluating GiveWell’s recommended charities is an important part of our work, and we take it seriously. It’s easy to talk about a commitment to evidence when the results are positive. It’s hard to maintain that commitment when the results are not. We’re excited to work with a group like Evidence Action that is committed to rigorous program evaluation and open discussion of the results of those evaluations. Its openness about these results has increased our confidence in Evidence Action as an organization. We look forward to seeing the results from the 2018 RCT in 2019.

Notes

[1] “At this early stage in analysis, we find no evidence that the program had an impact (positive or negative) on migration, caloric intake, food expenditure, or income.” Evidence Action, unpublished summary document, Page 1.

[2] The 2017 RCT studied a period from the fall of 2017 through early 2018.

[3] “This study has two main goals:

1. “A replication of previous findings showing positive impact of incentivized migration on seasonal migration, caloric intake, food and non-food expenditure, income, and food security. Our aim is to estimate impact of a scaled version of the No Lean Season program: intensifying program implementation within branches and expanding the provision of loans to all eligible households.”

Unpublished summary document, Page 1.

[4] “The second set of explanations focus on unintentional implementation changes caused by the change ineligibility, the vastly expanded scope of the program, or other factors. In the most recent round, it is possible that Migration Organizers (MOs) focused their efforts on those households who were most likely to migrate even without a loan to the exclusion of the target population households who need a loan to afford migration. Such behavior may have even been encouraged by the use of targets set by the NGO to manage implementation at such a large scale. We have implemented a qualitative survey to understand the incentives and actions of MOs last year, and are revising our instructions to avoid any possibility of this issue this year.” Evidence Action, unpublished summary document (with minor revision from Evidence Action), Page 11.

[5] “Most notably, the program was affected by severe flooding in many regions, and implementation was subsequently delayed as well. We are still evaluating whether these regions are the ones with the most diminished effects, although we lack the data in control areas to conduct an experimental comparison.” Evidence Action, unpublished summary document, Page 11-12.

[6] “It is possible that what we observe this year may be the true effect of the No Lean Season program when implemented at scale. This may be because conditions in rural Bangladesh have changed since the initial years of success, spillovers at scale cancel out any gains observed in small-scale pilots, or pilot villages were selected because they were most likely to be receptive to the program.” Evidence Action, unpublished summary document, Page 11.

[7] Evidence Action, “Interpretation of 2017 Results” deck and narrative (unpublished)

[8] “This adjustment is used to account for external validity concerns not accounted for elsewhere in the CEA.

“The default adjustment value of 80% is our best guess about the appropriate value, but it is not based on a formal calculation.

“The program at scale takes place in the same region with the same implementers (RDRS and Evidence Action) as the source of our key evidence for the intervention (the 2014 RCT). The program at scale differs in some aspects of implementation, particularly the inclusiveness of the eligibility criteria and the proportion of eligible households offered an incentive. In the 2014 RCT, the subsidy was a cash transfer rather than an interest-free loan, however the 2008 RCT found a similar effect regardless of whether the subsidy was a cash transfer or an interest-free loan.

“There is some evidence (from a 2013 RCT) suggesting that the program may be ineffective when the perceived risk of migrating increases for reasons such as labor strikes and violence. The researchers estimated that these are 1-in-10 year events.

“Additional discussion related to this parameter can be found at https://www.givewell.org/charities/no-lean-season#programdifferentfromRCTs.” 2018 GiveWell Cost-Effectiveness Model — Version 10, “Migration subsidies” tab, note on cell A19.

The post Update on No Lean Season’s top charity status appeared first on The GiveWell Blog.

At the end of 2017, we named Evidence Action’s No Lean Season one of GiveWell’s nine top charities. Now, GiveWell and Evidence Action agree that No Lean Season should not be a GiveWell top charity this year, and Evidence Action is not seeking additional funding to support its work at this time.

This post will discuss this decision in detail. In brief, we updated our assessment of No Lean Season, a program that provides loans to support seasonal migration, based on preliminary results Evidence Action began discussing with us in July from a study of the 2017 implementation of the program (hereinafter referred to as “2017 RCT”). These results suggested the program, as implemented in 2017, did not successfully induce migration. Taking this new information into account alongside previous studies of the program, we and Evidence Action do not believe No Lean Season meets our top charity criteria at this time.

Evidence Action’s post on this decision is here.

GiveWell’s mission is to identify and recommend charities that can most effectively use additional donations. While it may be disappointing for a top charity to be removed from our list of recommendations, we believe that adding and removing top charities from our list is an important part of our process. If our top charities list never changed, we would guess we were (a) acting too conservatively (i.e. not being open enough to adding new top charities), or (b) not being critical enough of groups once they’ve been added to our list (i.e. not being open enough to removing existing top charities).

We believe this decision speaks positively of Evidence Action and demonstrates our mutual commitment to updating our views based on new evidence. GiveWell has interacted with hundreds of organizations in our history, and very few have subjected their programs to a rigorous study in the way that Evidence Action did last year and, at smaller scale, in 2014. We’re excited to work with a group like Evidence Action that is committed to rigorous study and openness about results.

Summary

In this post, we will discuss:

• The history of GiveWell and No Lean Season. (More)
• How the 2017 RCT updated our views of No Lean Season. (More)
  • What did the 2017 RCT find? (More)
  • How did we interpret the RCT results? (More)
  • What does the future of No Lean Season look like? (More)
• Conclusion

GiveWell and No Lean Season

No Lean Season provides support for low-income agricultural workers in rural Bangladesh during the time of seasonal income and food insecurity (“lean season”). The program provides small, interest-free loans to support workers’ temporary migration to seek employment. No Lean Season is implemented by RDRS Bangladesh; Evidence Action provides strategic direction, conducts program monitoring, and provides technical assistance, among other functions. Evidence Action developed No Lean Season as part of its Beta portfolio, which is focused on prototyping and scaling cost-effective programs.

GiveWell began engaging with No Lean Season as a potential top charity in 2013, when we began to explore making an Incubation Grant to support its scale-up. We saw No Lean Season as a promising program that lacked the track record to be considered for a top charity recommendation at that time. We describe our initial interest in the program in a February 2017 blog post:

We approached Evidence Action in late 2013 to express our interest in supporting the creation of new GiveWell top charities.

In March 2014, Good Ventures made a $250,000 grant to Evidence Action to support the investigation and scale-up of promising programs.

Since then, Good Ventures has made three additional grants totaling approximately $2.7 million to support the program’s scale-up.

No Lean Season continued to test and scale their program with this and other support. We decided to recommend No Lean Season as a top charity in late 2017. We based our recommendation on three randomized controlled trials (RCTs) of the program. (We generally consider RCTs to be one of the strongest types of evidence available; you can read more about why we rely on RCTs here.)

Two of the RCTs (conducted in 2008 and 2014) indicated increased migration, income, and consumption for program participants. In the third RCT, which was conducted in 2013 and has not been published, the program is considered to have failed to induce migration, potentially due to political violence that year. We discuss the RCT evidence in greater depth in our intervention report on conditional subsidies for seasonal labor migration in northern Bangladesh.

Weighing the evidence, the cost of the program, and the potential impacts, we decided No Lean Season met our criteria to be named a top charity in November 2017. We summarized our reasoning in our blog post announcing our 2017 list of top charities, and noted the risks of this recommendation:

Several randomized controlled trials (RCTs) of subsidies to increase migration provide moderately strong evidence that such an intervention increases household income and consumption during the lean season. An additional RCT is ongoing. We estimate that No Lean Season is roughly five times as cost-effective as cash transfers (see our cost-effectiveness analysis).

Evidence Action has shared some details of its plans for monitoring No Lean Season in the future, but, as many of these plans have not been fully implemented, we have seen limited results. Therefore, there is some uncertainty as to whether No Lean Season will produce the data required to give us confidence that loans are appropriately targeted and reach their intended recipients in full; that recipients are not pressured into accepting loans; and that participants successfully migrate, find work, and are not exposed to major physical and other risks while migrating.

As indicated above, No Lean Season conducted an additional RCT to evaluate its program during the 2017 lean season (approximately September to December), the preliminary results of which indicate the program failed to induce migration. With the evidence from the 2017 RCT, the case for the program’s impact and cost-effectiveness looks weaker.

Our updated perspective on No Lean Season

The 2017 RCT was a key factor in the decision to remove No Lean Season from our top charities list. Below, we discuss:

• What the 2017 RCT found.
• Our interpretation of the RCT results.
• No Lean Season’s future.

What did the 2017 RCT find?

The 2017 RCT was a collaboration between Evidence Action, Innovations for Poverty Action, and researchers from Yale University, the London School of Economics, and the University of California, Davis. In a preliminary analysis shared with GiveWell in September 2018, the researchers did not find evidence for a negative or positive impact on migration, and found no statistically significant impact on income and consumption.[1]

However, the implementation of the program during the 2017[2] lean season and the evaluation of it differed from previous iterations. No Lean Season operated at a larger scale in the fall of 2017 than it had previously, offering loans to 158,155 households, compared with 16,268 households in 2016. Relative to earlier versions of the program, the program in 2017 involved (a) higher-intensity delivery of the intervention (offering loans to most eligible individuals) and (b) broader eligibility requirements (the eligibility rate in 2017 was 77 percent, compared with 49 percent in 2016).[3]

At this point, neither GiveWell, nor No Lean Season, nor the researchers feel we have a conclusive understanding of why the program failed to induce migration. However, No Lean Season and the researchers are exploring various hypotheses about what may explain the failure to induce migration, and they note that some suggestive evidence supports some hypotheses more than others. The researchers have posited several possibilities:

1. The way the program was targeted in 2017 was suboptimal. The Migration Organizers, who survey households for eligibility and offer and disburse loans (more detail here under “Migration Organizers”), may have focused their efforts on the individuals that were seen as most likely to migrate, rather than those who needed a loan to afford migration. The use of loan targets during implementation may have inadvertently incentivized this behavior.[4] If, for example, loan officers mostly made loans to people who would have migrated regardless of receiving a loan, this could have led to the lack of impact on migration found in the study.
2. The 2017 lean season was particularly bad for the program. The researchers note that severe flooding and associated implementation delays in some regions may have caused problems in 2017. The researchers plan to look more closely at the regions that experienced flooding, though they note that they don’t have the data necessary to make experimental comparisons.[5] In addition, a 2013 trial may have failed due to issues that were specific to the year of that trial, such as increased labor strikes.
3. There exists another (currently unknown) reason why this program won’t work at scale. Conditions in Bangladesh may have changed, negative spillovers (harmful impacts for individuals who did not receive loans) may cancel out gains, or pilot villages may have been strategically picked in earlier trials.[6]

The researchers are considering all of these possibilities. After considering various possible theories as well as some non-experimental data (including administrative data and data from a special-purpose survey of Migration Organizers who worked on the program in 2017), they feel that the ‘mistargeting’ theory is the most likely explanation and the explanation most consistent with the analysis.[7]

In scenario (1), No Lean Season may be able to identify and fix the problem. In scenario (2), GiveWell will need to update our estimate of the impact of the program to take into account the fact that periodic program failures due to external factors are more likely than we previously thought. In scenario (3), the program is unlikely to be effective in the future.

How did we interpret the RCT results?

We don’t know the extent to which each of the above explanations contributed to the study not finding an effect on migration.

We used the results of the 2017 RCT to update our cost-effectiveness estimate for the program. Cost-effectiveness estimates form arguably the most important single input into our decisions about whether or not to recommend charities (more on how GiveWell uses cost-effectiveness analyses here). When we calculate a program’s cost-effectiveness, we take many different factors into account, such as the administrative and program costs and the expected impact. We also make a number of educated guesses, such as the likelihood that a program’s impact in a new country will be similar to that in a country where it has previously worked. Below, we describe the mechanism by which the 2017 RCT result was incorporated into our model and how it changed our conclusion.

Prior to this year, we formed our view of No Lean Season based on the three small-scale RCTs mentioned above (conducted in 2008, 2013, and 2014). Each of these RCTs looked at a slightly different version of the program. We believed that the ‘high-intensity’ arm of the 2014 RCT was the version most likely to resemble the program at scale. We thus used the migration rate measured in this arm of the RCT as our starting point for calculating the program’s impact.

The high-intensity arm of the 2014 RCT also had the highest measured migration rate of the three RCTs we assessed, and so we wanted to give some consideration to the less-positive results found in the other two assessments. We applied a small, downward adjustment to the rate of induced migration observed in the 2014 high-intensity arm in our cost-effectiveness model; this was an educated guess, based on the information we had. Our best guess was that the program would lead, in expectation, to 80% of the induced migration seen in the 2014 high-intensity arm.[8]

Now, the preliminary 2017 RCT results show no significant impact on migration rates or incomes. Because this trial was large and very recent, we updated our expectations of the impact of the program substantially, and in a negative direction. Our best guess now is that the program will lead, in expectation, to 40% of the induced migration seen in the 2014 high-intensity arm. Holding other inputs constant, this adjustment reduces our estimate of No Lean Season’s cost-effectiveness by a factor of two.

This reduced cost-effectiveness, along with our updated qualitative picture of No Lean Season’s evidence of effectiveness, led to the decision to remove No Lean Season from our top charities list.

What does the future of No Lean Season look like?

Although they are not raising more funding at this time, No Lean Season has over two years’ worth of remaining funding. We understand that the organization has made changes to the program design in 2018 based on emerging interpretations of the 2017 results, and has collected additional data to evaluate some of the hypotheses which may explain those results (including, for example, a survey of Migration Organizers who worked on the 2017 program). They plan to subject the 2018 implementation round to an additional ‘RCT-at-scale,’ with a particular focus on reassessing the program’s effects on migration, income and consumption, as well as potential effects at migration destinations. They will continue to explore what may have caused the issue in the 2017 program at scale, and to see whether they can find a solution. If they do that, we’ll want to reassess the evidence and the costs to determine whether No Lean Season meets our bar for top charity status. Evidence Action believes we should have the necessary information to reassess starting in mid-2019, based on the results of the RCT conducted during the 2018 lean season and other analyses they perform.

Conclusion

This is the second time since 2011 that we have removed a top charity from our list (prior to 2011, our top charities list was fairly different from today; we made a big-picture shift in our priorities that year that led us to our more recent lists). The previous removal occurred in 2013, when we took the Against Malaria Foundation (AMF) off of our list because we didn’t believe it could absorb additional funding effectively in the near term. AMF was reinstated as a top charity in 2014.

The decision to remove a top charity is never easy. But continuously evaluating GiveWell’s recommended charities is an important part of our work, and we take it seriously. It’s easy to talk about a commitment to evidence when the results are positive. It’s hard to maintain that commitment when the results are not. We’re excited to work with a group like Evidence Action that is committed to rigorous program evaluation and open discussion of the results of those evaluations. Its openness about these results has increased our confidence in Evidence Action as an organization. We look forward to seeing the results from the 2018 RCT in 2019.

Notes

[1] “At this early stage in analysis, we find no evidence that the program had an impact (positive or negative) on migration, caloric intake, food expenditure, or income.” Evidence Action, unpublished summary document, Page 1.

[2] The 2017 RCT studied a period from the fall of 2017 through early 2018.

[3] “This study has two main goals:

1. “A replication of previous findings showing positive impact of incentivized migration on seasonal migration, caloric intake, food and non-food expenditure, income, and food security. Our aim is to estimate impact of a scaled version of the No Lean Season program: intensifying program implementation within branches and expanding the provision of loans to all eligible households.”

Unpublished summary document, Page 1.

[4] “The second set of explanations focus on unintentional implementation changes caused by the change ineligibility, the vastly expanded scope of the program, or other factors. In the most recent round, it is possible that Migration Organizers (MOs) focused their efforts on those households who were most likely to migrate even without a loan to the exclusion of the target population households who need a loan to afford migration. Such behavior may have even been encouraged by the use of targets set by the NGO to manage implementation at such a large scale. We have implemented a qualitative survey to understand the incentives and actions of MOs last year, and are revising our instructions to avoid any possibility of this issue this year.” Evidence Action, unpublished summary document (with minor revision from Evidence Action), Page 11.

[5] “Most notably, the program was affected by severe flooding in many regions, and implementation was subsequently delayed as well. We are still evaluating whether these regions are the ones with the most diminished effects, although we lack the data in control areas to conduct an experimental comparison.” Evidence Action, unpublished summary document, Page 11-12.

[6] “It is possible that what we observe this year may be the true effect of the No Lean Season program when implemented at scale. This may be because conditions in rural Bangladesh have changed since the initial years of success, spillovers at scale cancel out any gains observed in small-scale pilots, or pilot villages were selected because they were most likely to be receptive to the program.” Evidence Action, unpublished summary document, Page 11.

[7] Evidence Action, “Interpretation of 2017 Results” deck and narrative (unpublished)

[8] “This adjustment is used to account for external validity concerns not accounted for elsewhere in the CEA.

“The default adjustment value of 80% is our best guess about the appropriate value, but it is not based on a formal calculation.

“The program at scale takes place in the same region with the same implementers (RDRS and Evidence Action) as the source of our key evidence for the intervention (the 2014 RCT). The program at scale differs in some aspects of implementation, particularly the inclusiveness of the eligibility criteria and the proportion of eligible households offered an incentive. In the 2014 RCT, the subsidy was a cash transfer rather than an interest-free loan, however the 2008 RCT found a similar effect regardless of whether the subsidy was a cash transfer or an interest-free loan.

“There is some evidence (from a 2013 RCT) suggesting that the program may be ineffective when the perceived risk of migrating increases for reasons such as labor strikes and violence. The researchers estimated that these are 1-in-10 year events.

“Additional discussion related to this parameter can be found at https://www.givewell.org/charities/no-lean-season#programdifferentfromRCTs.” 2018 GiveWell Cost-Effectiveness Model — Version 10, “Migration subsidies” tab, note on cell A19.

The post Update on No Lean Season’s top charity status appeared first on The GiveWell Blog.

In July 2018, we recommended a $5.1 million grant to Evidence Action Beta to create a program dedicated to developing potential GiveWell top charities by prototyping, testing, and scaling programs which have the potential to be highly impactful and cost-effective.

This grant was made as part of GiveWell’s Incubation Grants program, which aims to support potential future GiveWell top charities and to help grow the pipeline of organizations we can consider for a recommendation. Funding for Incubation Grants comes from Good Ventures, a large foundation with which we work closely.

Summary

This post will discuss the following:

• Why Evidence Action Beta is promising. (More)
• Risks we see with this Incubation Grant. (More)
• Our plans for following Evidence Action Beta’s work going forward. (More)

Read More

The post A grant to Evidence Action Beta to prototype, test, and scale promising programs appeared first on The GiveWell Blog.

In July 2018, we recommended a $5.1 million grant to Evidence Action Beta to create a program dedicated to developing potential GiveWell top charities by prototyping, testing, and scaling programs which have the potential to be highly impactful and cost-effective.

This grant was made as part of GiveWell’s Incubation Grants program, which aims to support potential future GiveWell top charities and to help grow the pipeline of organizations we can consider for a recommendation. Funding for Incubation Grants comes from Good Ventures, a large foundation with which we work closely.

Summary

This post will discuss the following:

• Why Evidence Action Beta is promising. (More)
• Risks we see with this Incubation Grant. (More)
• Our plans for following Evidence Action Beta’s work going forward. (More)

Incubation Grant to Evidence Action Beta

We summarized our case for making this grant in a recently-published write-up:

A key part of GiveWell’s research process is trying to identify evidence-backed, cost-effective programs. GiveWell sometimes finds programs that seem potentially highly impactful based on academic research, but for which there is no obvious organizational partner that could scale up and test them. This grant will fund Evidence Action Beta to create … [an] incubator … focused on interventions that GiveWell and Evidence Action believe are promising but that lack existing organizations to scale them.

We have found that which program a charity works on is generally the most important factor in determining its overall cost-effectiveness. Through partnering with Evidence Action Beta to test programs that we think have the potential to be very cost-effective, … our hope is that programs tested and scaled up through this partnership may eventually become GiveWell top charities.

We believe this incubator has the potential to fill a major gap in the nonprofit world by providing a well-defined path for testing and potentially scaling … promising idea[s] for helping the global poor.

For full details on the grant activities and budget, see this page.

We believe that Evidence Action Beta is well-positioned to run this incubator because of its track record of scaling up cost-effective programs with high-quality monitoring. Evidence Action Beta’s parent organization, Evidence Action, leads two of our top charities (Deworm the World Initiative and No Lean Season) and one standout charity (Dispensers for Safe Water).

Modeling cost-effectiveness

In addition to the theoretical case for the grant outlined above, we also made explicit predictions and modeled the potential cost-effectiveness of this grant, so we could better consider it relative to other options. In this section, we provide more details on our process for estimating the grant’s cost-effectiveness.

The main path to impact we see with this grant is by creating new top charities which could use GiveWell-directed funds more cost-effectively than alternatives could.

This could occur:

1. if Evidence Action Beta incubates charities which are more cost-effective than our current top charities, or
2. if Evidence Action Beta incubates charities which are similarly cost-effective to our current top charities—in a scenario in which we have mostly filled our current top charities’ funding gaps. Right now, we believe our top charities can absorb significantly more funding than we expect to direct to them; this diminishes our view of the value of finding additional, similarly cost-effective opportunities. If our current top charities’ funding gaps were close to filled, we would place higher value on identifying additional room for more funding at a similarly cost-effective level.

This grant could also have an impact if it causes other, non-GiveWell funders to allocate resources to charities incubated by this grant. This incubator may create programs that GiveWell doesn’t direct funding to but others do. If these new opportunities are more cost-effective than what these funders would have otherwise supported, then this grant will have had a positive impact by causing funds to be spent more cost-effectively, even if GiveWell never recommends funding to the new programs directly.

We register forecasts for all Incubation Grants we make. We register these not because we are confident in them but because they help us clarify and communicate our expectation for the outcomes of the grant. Here, we forecast a 55% chance that Evidence Action Beta’s incubator leads to a new top charity by December 2023 that is 1-2x as cost-effective as the giving opportunity to which we would have otherwise directed those funds and a 30% chance that the grant does not lead to any new top charities by that time. (For more forecasts we made surrounding this grant, see here.)

We incorporated our forecasts as well as the potential impacts outlined above in our cost-effectiveness estimate for the grant: note that the potential upside coming from other funders is a particularly rough estimate which could change substantially with additional research.

Our best guess is that this grant is approximately ~9x as cost-effective as cash transfers, but we have spent limited time on this estimate and are highly uncertain about it. For context, we estimate that the average cost-effectiveness of our current top charities is between ~3x and ~12x as cost-effective as cash transfers.

Risks to the success of the grant

We do see risks to the success of this grant:

• Few programs may be more cost-effective than our current top charities, or our top charities may remain underfunded for a long time. If Evidence Action Beta fails to identify more cost-effective giving opportunities than GiveWell’s 2017 top charities, or if it only identifies similarly cost-effective giving opportunities while our current top charities remain underfunded, barring any major upside effects, this grant will have failed to make an impact.
• We expect this partnership with Evidence Action Beta to require a fair amount of senior staff capacity. If other means of identifying cost-effective giving opportunities, such as our work to evaluate policy opportunities, end up seeming more promising, this capacity may have been misused.

Going forward

This grant initiates a partnership with Evidence Action Beta toward which we might contribute substantial additional GiveWell Incubation Grant funding in the future. We plan to spend a fair amount of staff time on this ongoing partnership and follow this work closely.

We look forward to sharing updates and the results.

The post A grant to Evidence Action Beta to prototype, test, and scale promising programs appeared first on The GiveWell Blog.

In July 2018, we recommended a $5.1 million grant to Evidence Action Beta to create a program dedicated to developing potential GiveWell top charities by prototyping, testing, and scaling programs which have the potential to be highly impactful and cost-effective.

This grant was made as part of GiveWell’s Incubation Grants program, which aims to support potential future GiveWell top charities and to help grow the pipeline of organizations we can consider for a recommendation. Funding for Incubation Grants comes from Good Ventures, a large foundation with which we work closely.

Summary

This post will discuss the following:

• Why Evidence Action Beta is promising. (More)
• Risks we see with this Incubation Grant. (More)
• Our plans for following Evidence Action Beta’s work going forward. (More)

Incubation Grant to Evidence Action Beta

We summarized our case for making this grant in a recently-published write-up:

A key part of GiveWell’s research process is trying to identify evidence-backed, cost-effective programs. GiveWell sometimes finds programs that seem potentially highly impactful based on academic research, but for which there is no obvious organizational partner that could scale up and test them. This grant will fund Evidence Action Beta to create … [an] incubator … focused on interventions that GiveWell and Evidence Action believe are promising but that lack existing organizations to scale them.

We have found that which program a charity works on is generally the most important factor in determining its overall cost-effectiveness. Through partnering with Evidence Action Beta to test programs that we think have the potential to be very cost-effective, … our hope is that programs tested and scaled up through this partnership may eventually become GiveWell top charities.

We believe this incubator has the potential to fill a major gap in the nonprofit world by providing a well-defined path for testing and potentially scaling … promising idea[s] for helping the global poor.

For full details on the grant activities and budget, see this page.

We believe that Evidence Action Beta is well-positioned to run this incubator because of its track record of scaling up cost-effective programs with high-quality monitoring. Evidence Action Beta’s parent organization, Evidence Action, leads two of our top charities (Deworm the World Initiative and No Lean Season) and one standout charity (Dispensers for Safe Water).

Modeling cost-effectiveness

In addition to the theoretical case for the grant outlined above, we also made explicit predictions and modeled the potential cost-effectiveness of this grant, so we could better consider it relative to other options. In this section, we provide more details on our process for estimating the grant’s cost-effectiveness.

The main path to impact we see with this grant is by creating new top charities which could use GiveWell-directed funds more cost-effectively than alternatives could.

This could occur:

1. if Evidence Action Beta incubates charities which are more cost-effective than our current top charities, or
2. if Evidence Action Beta incubates charities which are similarly cost-effective to our current top charities—in a scenario in which we have mostly filled our current top charities’ funding gaps. Right now, we believe our top charities can absorb significantly more funding than we expect to direct to them; this diminishes our view of the value of finding additional, similarly cost-effective opportunities. If our current top charities’ funding gaps were close to filled, we would place higher value on identifying additional room for more funding at a similarly cost-effective level.

This grant could also have an impact if it causes other, non-GiveWell funders to allocate resources to charities incubated by this grant. This incubator may create programs that GiveWell doesn’t direct funding to but others do. If these new opportunities are more cost-effective than what these funders would have otherwise supported, then this grant will have had a positive impact by causing funds to be spent more cost-effectively, even if GiveWell never recommends funding to the new programs directly.

We register forecasts for all Incubation Grants we make. We register these not because we are confident in them but because they help us clarify and communicate our expectation for the outcomes of the grant. Here, we forecast a 55% chance that Evidence Action Beta’s incubator leads to a new top charity by December 2023 that is 1-2x as cost-effective as the giving opportunity to which we would have otherwise directed those funds and a 30% chance that the grant does not lead to any new top charities by that time. (For more forecasts we made surrounding this grant, see here.)

We incorporated our forecasts as well as the potential impacts outlined above in our cost-effectiveness estimate for the grant: note that the potential upside coming from other funders is a particularly rough estimate which could change substantially with additional research.

Our best guess is that this grant is approximately ~9x as cost-effective as cash transfers, but we have spent limited time on this estimate and are highly uncertain about it. For context, we estimate that the average cost-effectiveness of our current top charities is between ~3x and ~12x as cost-effective as cash transfers.

Risks to the success of the grant

We do see risks to the success of this grant:

• Few programs may be more cost-effective than our current top charities, or our top charities may remain underfunded for a long time. If Evidence Action Beta fails to identify more cost-effective giving opportunities than GiveWell’s 2017 top charities, or if it only identifies similarly cost-effective giving opportunities while our current top charities remain underfunded, barring any major upside effects, this grant will have failed to make an impact.
• We expect this partnership with Evidence Action Beta to require a fair amount of senior staff capacity. If other means of identifying cost-effective giving opportunities, such as our work to evaluate policy opportunities, end up seeming more promising, this capacity may have been misused.

Going forward

This grant initiates a partnership with Evidence Action Beta toward which we might contribute substantial additional GiveWell Incubation Grant funding in the future. We plan to spend a fair amount of staff time on this ongoing partnership and follow this work closely.

We look forward to sharing updates and the results.

The post A grant to Evidence Action Beta to prototype, test, and scale promising programs appeared first on The GiveWell Blog.

We’ve recently made a number of adjustments to improve our research process. Not all of them are easily visible outside of the organization.

This post is to highlight one of them: Publishing more frequent updates to our cost-effectiveness model throughout the year.

Summary

This post will explain:

• What changed in how we make updates to our cost-effectiveness model. (More)
• Why we made this change. (More)
• How to engage with updates to our model. (More)

What changed?

Last week, we published the ninth and tenth versions of our cost-effectiveness model in 2018. We made a number of updates to the newest versions of the model. They included accounting for reductions in malaria incidence for individuals who don’t receive seasonal malaria chemoprevention (SMC), the treatment one of our top charities distributes to prevent malaria, but who might benefit from living near other people receiving SMC (version 9) and the cost per deworming treatment delivered by another top charity, Sightsavers (version 10). These changes, and six others that were incorporated in the two latest versions, are described in our changelog.

Up until last year, we generally updated our cost-effectiveness model once or twice per year. However, as our model grew in complexity and we dedicated more research staff capacity to it, we decided that it would be beneficial to publish updates more regularly. We published our first in this series of more-frequent updates to our cost-effectiveness model in May 2017, as well as “release notes” (PDF) detailing the changes we made and the impact each had on our cost-effectiveness estimates.

We published five versions of our cost-effectiveness model in 2017. In 2018, we shifted from publishing PDF release notes to creating a “changelog“—a public page listing the changes we made to each version of the model, to be updated in tandem with the publication of each new version.

Internally, we moved toward having one staff member, Christian Smith, who is responsible for managing all changes to our cost-effectiveness model. He aims to publish a new version whenever there is a large, structurally complicated change to the model, or if there are several small and simple changes. Our internal process prioritizes being able to track how each change to the model moves the bottom line.

Changes we’ve published this year include updated inputs based on new research, such as the impact of insecticide resistance on the effectiveness of insecticide-treated nets; changes to inputs we include or exclude from the model altogether, such as removing short-term health benefits from deworming; and cosmetic changes to make the model easier to engage with, such as removing adjustments to account for the influence of GiveWell’s top charities on other actors from a particular tab.

Why we moved to this approach

Although it involves uncertainty, GiveWell’s cost-effectiveness model is a core piece of our research work and important input into our decisions about which charities to research and recommend. However, we believe it is challenging to engage with our model—to give a sense of the scale, our current model has 16 tabs, some of which use over 100 rows—and to keep up with changes we’ve made to the model over time.

Our hope is that publishing more frequent and transparent updates brings us closer in line to our goal of intense transparency and presenting a clear, vettable case for our recommendations to the public. It makes clearer the magnitude of any given change’s impact on our bottom line, and makes the evolution of the model over time easier to track. We also expect that it reduces the likelihood for errors, as fewer elements are being changed at any given time.

How to engage with updates to our model

We update our changelog, viewable here, when we publish a new version.

Going forward, we also plan to publish an announcement to our “Newly published GiveWell materials” email list when we do this. You can sign up to receive alerts from this email address here.

The post Publishing more frequent updates to our cost-effectiveness model appeared first on The GiveWell Blog.

We’ve recently made a number of adjustments to improve our research process. Not all of them are easily visible outside of the organization.

This post is to highlight one of them: Publishing more frequent updates to our cost-effectiveness model throughout the year.

Summary

This post will explain:

• What changed in how we make updates to our cost-effectiveness model. (More)
• Why we made this change. (More)
• How to engage with updates to our model. (More)

What changed?

Last week, we published the ninth and tenth versions of our cost-effectiveness model in 2018. We made a number of updates to the newest versions of the model. They included accounting for reductions in malaria incidence for individuals who don’t receive seasonal malaria chemoprevention (SMC), the treatment one of our top charities distributes to prevent malaria, but who might benefit from living near other people receiving SMC (version 9) and the cost per deworming treatment delivered by another top charity, Sightsavers (version 10). These changes, and six others that were incorporated in the two latest versions, are described in our changelog.

Up until last year, we generally updated our cost-effectiveness model once or twice per year. However, as our model grew in complexity and we dedicated more research staff capacity to it, we decided that it would be beneficial to publish updates more regularly. We published our first in this series of more-frequent updates to our cost-effectiveness model in May 2017, as well as “release notes” (PDF) detailing the changes we made and the impact each had on our cost-effectiveness estimates.

We published five versions of our cost-effectiveness model in 2017. In 2018, we shifted from publishing PDF release notes to creating a “changelog“—a public page listing the changes we made to each version of the model, to be updated in tandem with the publication of each new version.

Internally, we moved toward having one staff member, Christian Smith, who is responsible for managing all changes to our cost-effectiveness model. He aims to publish a new version whenever there is a large, structurally complicated change to the model, or if there are several small and simple changes. Our internal process prioritizes being able to track how each change to the model moves the bottom line.

Changes we’ve published this year include updated inputs based on new research, such as the impact of insecticide resistance on the effectiveness of insecticide-treated nets; changes to inputs we include or exclude from the model altogether, such as removing short-term health benefits from deworming; and cosmetic changes to make the model easier to engage with, such as removing adjustments to account for the influence of GiveWell’s top charities on other actors from a particular tab.

Why we moved to this approach

Although it involves uncertainty, GiveWell’s cost-effectiveness model is a core piece of our research work and important input into our decisions about which charities to research and recommend. However, we believe it is challenging to engage with our model—to give a sense of the scale, our current model has 16 tabs, some of which use over 100 rows—and to keep up with changes we’ve made to the model over time.

Our hope is that publishing more frequent and transparent updates brings us closer in line to our goal of intense transparency and presenting a clear, vettable case for our recommendations to the public. It makes clearer the magnitude of any given change’s impact on our bottom line, and makes the evolution of the model over time easier to track. We also expect that it reduces the likelihood for errors, as fewer elements are being changed at any given time.

How to engage with updates to our model

We update our changelog, viewable here, when we publish a new version.

Going forward, we also plan to publish an announcement to our “Newly published GiveWell materials” email list when we do this. You can sign up to receive alerts from this email address here.

The post Publishing more frequent updates to our cost-effectiveness model appeared first on The GiveWell Blog.

Our goal with hosting quarterly open threads is to give blog readers an opportunity to publicly raise comments or questions about GiveWell or related topics (in the comments section below). As always, you’re also welcome to email us at info@givewell.org or to request a call with GiveWell staff if you have feedback or questions you’d prefer to discuss privately. We’ll try to respond promptly to questions or comments.

You can view our June 2018 open thread here.

The post September 2018 open thread appeared first on The GiveWell Blog.

Our goal with hosting quarterly open threads is to give blog readers an opportunity to publicly raise comments or questions about GiveWell or related topics (in the comments section below). As always, you’re also welcome to email us at info@givewell.org or to request a call with GiveWell staff if you have feedback or questions you’d prefer to discuss privately. We’ll try to respond promptly to questions or comments.

You can view our June 2018 open thread here.

The post September 2018 open thread appeared first on The GiveWell Blog.

In April to June 2018, we received $1.2 million in funding for making grants at our discretion. In addition, GiveWell’s Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. In this post we discuss:

• The decision to allocate the $4.1 million to the Against Malaria Foundation (AMF) (70 percent) and the Schistosomiasis Control Initiative (SCI) (30 percent).
• Our recommendation that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we continue to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact.
• Why we have allocated unrestricted funds to making grants to recommended charities.

Allocation of discretionary funds

The allocation of 70 percent of the funds to AMF and 30 percent to SCI follows the recommendation we have made, and continue to make, to donors. For more discussion on this allocation, see our blog post about allocating discretionary funds from the fourth quarter of 2017.

We ask each top charity to provide details of how they will use additional funding each year, as part of our process to update our “room for more funding” summary for each top charity. This year, we have asked for this information by the end of July. We also ask each of our top charities to let us know if they encounter unexpected funding gaps at other times of year. We have not learned of new funding gaps in the last quarter.

What is our recommendation to donors?

We continue to recommend that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we are continuing to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact. The reasons for this recommendation are the same as in our Q4 2017 post on allocating discretionary funding.

We will complete a full analysis of our top charities’ funding gaps and cost-effectiveness by November and expect to update our recommendation to donors at that time.

Why we have allocated unrestricted funds to making grants to recommended charities

In June, GiveWell’s Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. We generally use unrestricted funds to support GiveWell’s operating costs. The decision was made to grant out some of the unrestricted funds we hold in accordance with two policies:

• Our “excess assets” policy specifies that once we surpass a certain level of unrestricted assets, we grant out the excess rather than continue to hold it ourselves. We reviewed our unrestricted asset holdings and projected revenue and expenses for 2018-2020 and concluded that we held $1.8 million more than was required to give us a stable, predictable financial situation (details of how this rule is applied are at the previous link). The Board voted to irrevocably restrict this amount to making grants to recommended charities. Note that we continue to need ongoing donor support for our operations. This decision incorporates our projections for future donations.
• In order to limit the risks of relying too heavily on any single source of revenue, we cap the amount of funding that we will use from one source to support our operating costs at 20% of our projected annual expenses. In early 2018, we received a donation of $2.1 million in unrestricted funds. Our operating expense budget for 2018 is $4.9 million. Therefore, the Board voted to retain $1.0 million to support operating costs in 2018 and irrevocably restrict $1.1 million to making grants to recommended charities.

The post Allocation of discretionary funds from Q2 2018 appeared first on The GiveWell Blog.

In April to June 2018, we received $1.2 million in funding for making grants at our discretion. In addition, GiveWell's Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. In this post we discuss:

• The decision to allocate the $4.1 million to the Against Malaria Foundation (AMF) (70 percent) and the Schistosomiasis Control Initiative (SCI) (30 percent).
• Our recommendation that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we continue to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact.
• Why we have allocated unrestricted funds to making grants to recommended charities.

Read More

The post Allocation of discretionary funds from Q2 2018 appeared first on The GiveWell Blog.

In April to June 2018, we received $1.2 million in funding for making grants at our discretion. In addition, GiveWell's Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. In this post we discuss:

• The decision to allocate the $4.1 million to the Against Malaria Foundation (AMF) (70 percent) and the Schistosomiasis Control Initiative (SCI) (30 percent).
• Our recommendation that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we continue to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact.
• Why we have allocated unrestricted funds to making grants to recommended charities.

Read More

The post Allocation of discretionary funds from Q2 2018 appeared first on The GiveWell Blog.

In April to June 2018, we received $1.2 million in funding for making grants at our discretion. In addition, GiveWell’s Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. In this post we discuss:

• The decision to allocate the $4.1 million to the Against Malaria Foundation (AMF) (70 percent) and the Schistosomiasis Control Initiative (SCI) (30 percent).
• Our recommendation that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we continue to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact.
• Why we have allocated unrestricted funds to making grants to recommended charities.

Allocation of discretionary funds

The allocation of 70 percent of the funds to AMF and 30 percent to SCI follows the recommendation we have made, and continue to make, to donors. For more discussion on this allocation, see our blog post about allocating discretionary funds from the fourth quarter of 2017.

We ask each top charity to provide details of how they will use additional funding each year, as part of our process to update our “room for more funding” summary for each top charity. This year, we have asked for this information by the end of July. We also ask each of our top charities to let us know if they encounter unexpected funding gaps at other times of year. We have not learned of new funding gaps in the last quarter.

What is our recommendation to donors?

We continue to recommend that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we are continuing to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact. The reasons for this recommendation are the same as in our Q4 2017 post on allocating discretionary funding.

We will complete a full analysis of our top charities’ funding gaps and cost-effectiveness by November and expect to update our recommendation to donors at that time.

Why we have allocated unrestricted funds to making grants to recommended charities

In June, GiveWell’s Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. We generally use unrestricted funds to support GiveWell’s operating costs. The decision was made to grant out some of the unrestricted funds we hold in accordance with two policies:

• Our “excess assets” policy specifies that once we surpass a certain level of unrestricted assets, we grant out the excess rather than continue to hold it ourselves. We reviewed our unrestricted asset holdings and projected revenue and expenses for 2018-2020 and concluded that we held $1.8 million more than was required to give us a stable, predictable financial situation (details of how this rule is applied are at the previous link). The Board voted to irrevocably restrict this amount to making grants to recommended charities. Note that we continue to need ongoing donor support for our operations. This decision incorporates our projections for future donations.
• In order to limit the risks of relying too heavily on any single source of revenue, we cap the amount of funding that we will use from one source to support our operating costs at 20% of our projected annual expenses. In early 2018, we received a donation of $2.1 million in unrestricted funds. Our operating expense budget for 2018 is $4.9 million. Therefore, the Board voted to retain $1.0 million to support operating costs in 2018 and irrevocably restrict $1.1 million to making grants to recommended charities.

The post Allocation of discretionary funds from Q2 2018 appeared first on The GiveWell Blog.

In April to June 2018, we received $1.2 million in funding for making grants at our discretion. In addition, GiveWell’s Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. In this post we discuss:

• The decision to allocate the $4.1 million to the Against Malaria Foundation (AMF) (70 percent) and the Schistosomiasis Control Initiative (SCI) (30 percent).
• Our recommendation that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we continue to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact.
• Why we have allocated unrestricted funds to making grants to recommended charities.

Allocation of discretionary funds

The allocation of 70 percent of the funds to AMF and 30 percent to SCI follows the recommendation we have made, and continue to make, to donors. For more discussion on this allocation, see our blog post about allocating discretionary funds from the fourth quarter of 2017.

We ask each top charity to provide details of how they will use additional funding each year, as part of our process to update our “room for more funding” summary for each top charity. This year, we have asked for this information by the end of July. We also ask each of our top charities to let us know if they encounter unexpected funding gaps at other times of year. We have not learned of new funding gaps in the last quarter.

What is our recommendation to donors?

We continue to recommend that donors give to GiveWell for granting to top charities at our discretion so that we can direct the funding to the top charity or charities with the most pressing funding need. For donors who prefer to give directly to our top charities, we are continuing to recommend giving 70 percent of your donation to AMF and 30 percent to SCI to maximize your impact. The reasons for this recommendation are the same as in our Q4 2017 post on allocating discretionary funding.

We will complete a full analysis of our top charities’ funding gaps and cost-effectiveness by November and expect to update our recommendation to donors at that time.

Why we have allocated unrestricted funds to making grants to recommended charities

In June, GiveWell’s Board of Directors voted to allocate $2.9 million in unrestricted funds to making grants to recommended charities. We generally use unrestricted funds to support GiveWell’s operating costs. The decision was made to grant out some of the unrestricted funds we hold in accordance with two policies:

• Our “excess assets” policy specifies that once we surpass a certain level of unrestricted assets, we grant out the excess rather than continue to hold it ourselves. We reviewed our unrestricted asset holdings and projected revenue and expenses for 2018-2020 and concluded that we held $1.8 million more than was required to give us a stable, predictable financial situation (details of how this rule is applied are at the previous link). The Board voted to irrevocably restrict this amount to making grants to recommended charities. Note that we continue to need ongoing donor support for our operations. This decision incorporates our projections for future donations.
• In order to limit the risks of relying too heavily on any single source of revenue, we cap the amount of funding that we will use from one source to support our operating costs at 20% of our projected annual expenses. In early 2018, we received a donation of $2.1 million in unrestricted funds. Our operating expense budget for 2018 is $4.9 million. Therefore, the Board voted to retain $1.0 million to support operating costs in 2018 and irrevocably restrict $1.1 million to making grants to recommended charities.

The post Allocation of discretionary funds from Q2 2018 appeared first on The GiveWell Blog.

Summary

We recently completed a small project to determine whether using subnational baseline malaria mortality estimates would make a difference to our estimates of the cost-effectiveness of two of our top charities, the Against Malaria Foundation and Malaria Consortium. We ultimately decided not to include these adjustments because they added complexity to our models and would require frequent updating, while only making a small difference (a 3-4% improvement) to our bottom line.

Though this post is on a fairly narrow topic, we believe this example illustrates the principles we use to make decisions about what to include in our cost-effectiveness model.

Background

Two of our top charities—the Against Malaria Foundation (AMF) and Malaria Consortium’s seasonal malaria chemoprevention program—implement programs to prevent malaria, a leading killer of people in low- and middle-income countries.

One of the core reasons we recommend AMF and Malaria Consortium is their cost-effectiveness: how much impact they have (e.g., cases of malaria prevented, malaria deaths averted) with the funds they receive. Our estimates of charities’ cost-effectiveness isn’t just helpful to us in determining which charities should be GiveWell top charities; we also rely on these estimates to guide our decisions about how to allocate funding between our top charities.

Our cost-effectiveness estimates for AMF and Malaria Consortium use country-wide data on malaria mortality and malaria incidence in the places that both organizations work.1In both cases, we rely on reports by Cochrane, an organization that produces systematic reviews and other synthesized research to inform decision-makers. For AMF, we use a decline in all-cause mortality, because the Cochrane review of anti-malarial bed net distributions reports the effect in terms of a reduction in all-cause mortality. For Malaria Consortium, we use a decline in malaria mortality (proxied by a decline in malaria incidence), as the Cochrane review of seasonal malaria chemoprevention reports the effect in terms of a reduction in malaria incidence, but not all-cause mortality. See our cost-effectiveness analysis for more details. jQuery("#footnote_plugin_tooltip_1").tooltip({ tip: "#footnote_plugin_tooltip_text_1", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); However, neither organization serves a whole country—rather, they operate in sub-national regions—so the use of country-level estimates could cause us to either underestimate or overestimate their cost-effectiveness. If, for example, these programs are focused in the areas of the country with the highest malaria burden, using the average burden for the country would lead us to underestimate their cost-effectiveness. So, we completed a project to determine how much of an impact using subnational estimates would have, to consider whether we ought to incorporate this information into our cost-effectiveness analysis.

How we estimated the impact of subnational malaria incidence

AMF distributes insecticide-treated nets to prevent malaria; Malaria Consortium’s seasonal malaria chemoprevention (SMC) program provides preventive anti-malarial drugs. We used estimates of subnational malaria incidence from the Malaria Atlas Project (MAP) to see if regions covered by nets or eligible for SMC had higher or lower incidence than the average in the country in which they are located.2We assume that the regional distribution of malaria incidence is a reasonable proxy for the regional distribution of malaria mortality. jQuery("#footnote_plugin_tooltip_2").tooltip({ tip: "#footnote_plugin_tooltip_text_2", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

We focused on all areas covered by nets or eligible for SMC (rather than those covered by our top charities, specifically) for two reasons:

1. Our understanding is that when our top charities contribute resources to a country’s net distribution or SMC programs, the marginal region covered by these additional resources is not necessarily the same as the region to which these resources are assigned (because these resources are fungible with other resources within the national programs).3A limitation of this analysis is it does not account for the possibility that AMF and Malaria Consortium are causing locations that are higher priority or lower priority than the average location already covered by nets or eligible for SMC to be covered on the margin. We do not explicitly include estimates of the marginal region funded in our cost-effectiveness analysis because we often have limited information about which regions would be covered with marginal additional funds. jQuery("#footnote_plugin_tooltip_3").tooltip({ tip: "#footnote_plugin_tooltip_text_3", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });
2. Our aim is to estimate the cost-effectiveness of funds donated to these organizations in the future. The subnational region where AMF has worked in the past has not historically been a good indicator of the region where it will work in future.

Results for net distributions in countries where AMF works

We looked at geographical variation in malaria incidence in countries where AMF works, weighting each region by the number of nets it currently receives.4We assume that where nets have been delivered in the past is a good proxy for where new nets will be delivered in the future. The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_4").tooltip({ tip: "#footnote_plugin_tooltip_text_4", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

The average net delivered in the countries in which AMF works is hung in an area with 0-9% higher malaria incidence than the average in that country, and the weighted average adjustment to AMF’s cost-effectiveness would be 3% (in other words, AMF becomes 3% more cost-effective if we incorporate subnational estimates).5See Cell J114. We did not include Papua New Guinea (where AMF funds some nets) in this analysis, as MAP only covers countries in Africa. jQuery("#footnote_plugin_tooltip_5").tooltip({ tip: "#footnote_plugin_tooltip_text_5", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

Country Adjustment Zambia +9% Uganda +4% Ghana +4% Democratic Republic of the Congo +1% Togo +1% Malawi +0% Results for SMC in countries where Malaria Consortium works

We looked at six countries comprising >95% of Malaria Consortium’s SMC spending and compared malaria incidence in districts eligible for SMC with the country-wide average.6“The suitability of an area for SMC is determined by the seasonal pattern of rainfall, malaria transmission and the burden of malaria. SMC is recommended for deployment in areas: (i) where more than 60% of the annual incidence of malaria occurs within 4 months (ii) where there are measures of disease burden consistent with a high burden of malaria in children (incidence ≥ 10 cases of malaria among every 100 children during the transmission season) (iii) where SP and AQ [the drugs used to treat children] retain their antimalarial efficacy.” WHO SMC field guide (2013), Pg 8. jQuery("#footnote_plugin_tooltip_6").tooltip({ tip: "#footnote_plugin_tooltip_text_6", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });7The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_7").tooltip({ tip: "#footnote_plugin_tooltip_text_7", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

The average region eligible for SMC in countries where Malaria Consortium works has -2% to 17% higher malaria incidence than the average in that country. The weighted average adjustment to Malaria Consortium’s cost-effectiveness would be 4%.8See Cell C126. jQuery("#footnote_plugin_tooltip_8").tooltip({ tip: "#footnote_plugin_tooltip_text_8", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

Country Adjustment Commentary Guinea +17% Conakry, the capital, is ineligible for SMC and has low incidence. Nigeria +12% SMC appears to be targeted in the north, where malaria incidence is slightly higher. Niger +2% The majority of the population is either covered or planned to be covered from 2019. Burkina Faso 0% All districts are eligible. Mali 0% All districts are eligible. Chad -2% The four regions with very low malaria incidence (Borkou, Tibesti, Ennedi Est and Ouest) aren’t eligible for SMC, but are sparsely populated. What we concluded

We decided not to include these adjustments in our cost-effectiveness analysis because they increased complexity, without substantially affecting the bottom line.

When we decide whether to include adjustments in our model in general, we use a framework that first takes our best guess of the likely effect size and then rates each of the remaining question on a three-point scale.

Score9We use these scores as a qualitative guide to help us think through what to include in our cost-effectiveness analysis. You can see the rubric we use to assign scores in this spreadsheet. jQuery("#footnote_plugin_tooltip_9").tooltip({ tip: "#footnote_plugin_tooltip_text_9", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); Commentary Best guess of effect size 3-4% Can it be objectively justified? 3/3 While we have not investigated the MAP data in detail, we would guess that after further investigation, we would conclude it provides a reasonable approximation of subnational malaria incidence.10You can read more about MAP’s methodology in this paper. jQuery("#footnote_plugin_tooltip_10").tooltip({ tip: "#footnote_plugin_tooltip_text_10", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); How easily can it be modelled? 3/3 The methodology is clear and simple. Is it consistent with our other cost-effectiveness analyses? 2/3 We could include subnational adjustments for both of our top charities that implement malaria-prevention programs, but we believe it is unlikely there would be sufficient data to do the same for prevalence of worms or vitamin A deficiency (the focus of five of our other seven top charities).

Even though these adjustments can be objectively justified and are fairly easy to model, the bottom-line difference they make to our cost-effectiveness estimates is insufficient to warrant the (moderate) increase in the complexity of our models. These adjustments would also introduce an inconsistency between our methodologies for top charities. As a result, we are not planning to incorporate subnational adjustments at this time.

When would we revisit this conclusion?

We will revisit using subnational malaria mortality estimates if AMF or Malaria Consortium start working in countries where it would make a large difference to the bottom line. We would include subnational adjustments if AMF contributed nets in any of these countries: Djibouti (+500% adjustment), South Africa (+259%), and Swaziland (+126%), where malaria is endemic in some parts of the country but not others. We would also consider subnational adjustments if AMF contributed nets in Namibia (+25%), Kenya (+23%), Madagascar (+14%), or Rwanda (+10%).11The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_11").tooltip({ tip: "#footnote_plugin_tooltip_text_11", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

We will investigate whether subnational adjustments would make a substantial difference if Malaria Consortium enters additional countries; at this time, we do not have details on which regions are eligible for SMC in countries in which Malaria Consortium is not currently operating.12We have not yet prioritized getting details on which regions are eligible for SMC in countries in which Malaria Consortium does not currently work, as this would likely impose a substantial time cost on Malaria Consortium. jQuery("#footnote_plugin_tooltip_12").tooltip({ tip: "#footnote_plugin_tooltip_text_12", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

You can read the internal emails discussing our decision process here.

Notes   [ + ]

1. ↑ In both cases, we rely on reports by Cochrane, an organization that produces systematic reviews and other synthesized research to inform decision-makers. For AMF, we use a decline in all-cause mortality, because the Cochrane review of anti-malarial bed net distributions reports the effect in terms of a reduction in all-cause mortality. For Malaria Consortium, we use a decline in malaria mortality (proxied by a decline in malaria incidence), as the Cochrane review of seasonal malaria chemoprevention reports the effect in terms of a reduction in malaria incidence, but not all-cause mortality. See our cost-effectiveness analysis for more details. 2. ↑ We assume that the regional distribution of malaria incidence is a reasonable proxy for the regional distribution of malaria mortality. 3. ↑ A limitation of this analysis is it does not account for the possibility that AMF and Malaria Consortium are causing locations that are higher priority or lower priority than the average location already covered by nets or eligible for SMC to be covered on the margin. We do not explicitly include estimates of the marginal region funded in our cost-effectiveness analysis because we often have limited information about which regions would be covered with marginal additional funds. 4. ↑ We assume that where nets have been delivered in the past is a good proxy for where new nets will be delivered in the future. The data and calculations are in this spreadsheet. 5. ↑ See Cell J114. We did not include Papua New Guinea (where AMF funds some nets) in this analysis, as MAP only covers countries in Africa. 6. ↑ “The suitability of an area for SMC is determined by the seasonal pattern of rainfall, malaria transmission and the burden of malaria. SMC is recommended for deployment in areas: (i) where more than 60% of the annual incidence of malaria occurs within 4 months (ii) where there are measures of disease burden consistent with a high burden of malaria in children (incidence ≥ 10 cases of malaria among every 100 children during the transmission season) (iii) where SP and AQ [the drugs used to treat children] retain their antimalarial efficacy.” WHO SMC field guide (2013), Pg 8. 7, 11. ↑ The data and calculations are in this spreadsheet. 8. ↑ See Cell C126. 9. ↑ We use these scores as a qualitative guide to help us think through what to include in our cost-effectiveness analysis. You can see the rubric we use to assign scores in this spreadsheet. 10. ↑ You can read more about MAP’s methodology in this paper. 12. ↑ We have not yet prioritized getting details on which regions are eligible for SMC in countries in which Malaria Consortium does not currently work, as this would likely impose a substantial time cost on Malaria Consortium. function footnote_expand_reference_container() { jQuery("#footnote_references_container").show(); jQuery("#footnote_reference_container_collapse_button").text("-"); } function footnote_collapse_reference_container() { jQuery("#footnote_references_container").hide(); jQuery("#footnote_reference_container_collapse_button").text("+"); } function footnote_expand_collapse_reference_container() { if (jQuery("#footnote_references_container").is(":hidden")) { footnote_expand_reference_container(); } else { footnote_collapse_reference_container(); } } function footnote_moveToAnchor(p_str_TargetID) { footnote_expand_reference_container(); var l_obj_Target = jQuery("#" + p_str_TargetID); if(l_obj_Target.length) { jQuery('html, body').animate({ scrollTop: l_obj_Target.offset().top - window.innerHeight/2 }, 1000); } }

The post Why we don’t use subnational malaria mortality estimates in our cost-effectiveness models appeared first on The GiveWell Blog.

Summary

We recently completed a small project to determine whether using subnational baseline malaria mortality estimates would make a difference to our estimates of the cost-effectiveness of two of our top charities, the Against Malaria Foundation and Malaria Consortium. We ultimately decided not to include these adjustments because they added complexity to our models and would require frequent updating, while only making a small difference (a 3-4% improvement) to our bottom line.

Though this post is on a fairly narrow topic, we believe this example illustrates the principles we use to make decisions about what to include in our cost-effectiveness model.

Background

Two of our top charities—the Against Malaria Foundation (AMF) and Malaria Consortium’s seasonal malaria chemoprevention program—implement programs to prevent malaria, a leading killer of people in low- and middle-income countries.

One of the core reasons we recommend AMF and Malaria Consortium is their cost-effectiveness: how much impact they have (e.g., cases of malaria prevented, malaria deaths averted) with the funds they receive. Our estimates of charities’ cost-effectiveness isn’t just helpful to us in determining which charities should be GiveWell top charities; we also rely on these estimates to guide our decisions about how to allocate funding between our top charities.

Our cost-effectiveness estimates for AMF and Malaria Consortium use country-wide data on malaria mortality and malaria incidence in the places that both organizations work.1In both cases, we rely on reports by Cochrane, an organization that produces systematic reviews and other synthesized research to inform decision-makers. For AMF, we use a decline in all-cause mortality, because the Cochrane review of anti-malarial bed net distributions reports the effect in terms of a reduction in all-cause mortality. For Malaria Consortium, we use a decline in malaria mortality (proxied by a decline in malaria incidence), as the Cochrane review of seasonal malaria chemoprevention reports the effect in terms of a reduction in malaria incidence, but not all-cause mortality. See our cost-effectiveness analysis for more details. jQuery("#footnote_plugin_tooltip_1").tooltip({ tip: "#footnote_plugin_tooltip_text_1", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); However, neither organization serves a whole country—rather, they operate in sub-national regions—so the use of country-level estimates could cause us to either underestimate or overestimate their cost-effectiveness. If, for example, these programs are focused in the areas of the country with the highest malaria burden, using the average burden for the country would lead us to underestimate their cost-effectiveness. So, we completed a project to determine how much of an impact using subnational estimates would have, to consider whether we ought to incorporate this information into our cost-effectiveness analysis.

How we estimated the impact of subnational malaria incidence

AMF distributes insecticide-treated nets to prevent malaria; Malaria Consortium’s seasonal malaria chemoprevention (SMC) program provides preventive anti-malarial drugs. We used estimates of subnational malaria incidence from the Malaria Atlas Project (MAP) to see if regions covered by nets or eligible for SMC had higher or lower incidence than the average in the country in which they are located.2We assume that the regional distribution of malaria incidence is a reasonable proxy for the regional distribution of malaria mortality. jQuery("#footnote_plugin_tooltip_2").tooltip({ tip: "#footnote_plugin_tooltip_text_2", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

We focused on all areas covered by nets or eligible for SMC (rather than those covered by our top charities, specifically) for two reasons:

1. Our understanding is that when our top charities contribute resources to a country’s net distribution or SMC programs, the marginal region covered by these additional resources is not necessarily the same as the region to which these resources are assigned (because these resources are fungible with other resources within the national programs).3A limitation of this analysis is it does not account for the possibility that AMF and Malaria Consortium are causing locations that are higher priority or lower priority than the average location already covered by nets or eligible for SMC to be covered on the margin. We do not explicitly include estimates of the marginal region funded in our cost-effectiveness analysis because we often have limited information about which regions would be covered with marginal additional funds. jQuery("#footnote_plugin_tooltip_3").tooltip({ tip: "#footnote_plugin_tooltip_text_3", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });
2. Our aim is to estimate the cost-effectiveness of funds donated to these organizations in the future. The subnational region where AMF has worked in the past has not historically been a good indicator of the region where it will work in future.

Results for net distributions in countries where AMF works

We looked at geographical variation in malaria incidence in countries where AMF works, weighting each region by the number of nets it currently receives.4We assume that where nets have been delivered in the past is a good proxy for where new nets will be delivered in the future. The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_4").tooltip({ tip: "#footnote_plugin_tooltip_text_4", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

The average net delivered in the countries in which AMF works is hung in an area with 0-9% higher malaria incidence than the average in that country, and the weighted average adjustment to AMF’s cost-effectiveness would be 3% (in other words, AMF becomes 3% more cost-effective if we incorporate subnational estimates).5See Cell J114. We did not include Papua New Guinea (where AMF funds some nets) in this analysis, as MAP only covers countries in Africa. jQuery("#footnote_plugin_tooltip_5").tooltip({ tip: "#footnote_plugin_tooltip_text_5", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

Country Adjustment Zambia +9% Uganda +4% Ghana +4% Democratic Republic of the Congo +1% Togo +1% Malawi +0% Results for SMC in countries where Malaria Consortium works

We looked at six countries comprising >95% of Malaria Consortium’s SMC spending and compared malaria incidence in districts eligible for SMC with the country-wide average.6“The suitability of an area for SMC is determined by the seasonal pattern of rainfall, malaria transmission and the burden of malaria. SMC is recommended for deployment in areas: (i) where more than 60% of the annual incidence of malaria occurs within 4 months (ii) where there are measures of disease burden consistent with a high burden of malaria in children (incidence ≥ 10 cases of malaria among every 100 children during the transmission season) (iii) where SP and AQ [the drugs used to treat children] retain their antimalarial efficacy.” WHO SMC field guide (2013), Pg 8. jQuery("#footnote_plugin_tooltip_6").tooltip({ tip: "#footnote_plugin_tooltip_text_6", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });7The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_7").tooltip({ tip: "#footnote_plugin_tooltip_text_7", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

The average region eligible for SMC in countries where Malaria Consortium works has -2% to 17% higher malaria incidence than the average in that country. The weighted average adjustment to Malaria Consortium’s cost-effectiveness would be 4%.8See Cell C126. jQuery("#footnote_plugin_tooltip_8").tooltip({ tip: "#footnote_plugin_tooltip_text_8", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

Country Adjustment Commentary Guinea +17% Conakry, the capital, is ineligible for SMC and has low incidence. Nigeria +12% SMC appears to be targeted in the north, where malaria incidence is slightly higher. Niger +2% The majority of the population is either covered or planned to be covered from 2019. Burkina Faso 0% All districts are eligible. Mali 0% All districts are eligible. Chad -2% The four regions with very low malaria incidence (Borkou, Tibesti, Ennedi Est and Ouest) aren’t eligible for SMC, but are sparsely populated. What we concluded

We decided not to include these adjustments in our cost-effectiveness analysis because they increased complexity, without substantially affecting the bottom line.

When we decide whether to include adjustments in our model in general, we use a framework that first takes our best guess of the likely effect size and then rates each of the remaining question on a three-point scale.

Score9We use these scores as a qualitative guide to help us think through what to include in our cost-effectiveness analysis. You can see the rubric we use to assign scores in this spreadsheet. jQuery("#footnote_plugin_tooltip_9").tooltip({ tip: "#footnote_plugin_tooltip_text_9", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); Commentary Best guess of effect size 3-4% Can it be objectively justified? 3/3 While we have not investigated the MAP data in detail, we would guess that after further investigation, we would conclude it provides a reasonable approximation of subnational malaria incidence.10You can read more about MAP’s methodology in this paper. jQuery("#footnote_plugin_tooltip_10").tooltip({ tip: "#footnote_plugin_tooltip_text_10", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); How easily can it be modelled? 3/3 The methodology is clear and simple. Is it consistent with our other cost-effectiveness analyses? 2/3 We could include subnational adjustments for both of our top charities that implement malaria-prevention programs, but we believe it is unlikely there would be sufficient data to do the same for prevalence of worms or vitamin A deficiency (the focus of five of our other seven top charities).

Even though these adjustments can be objectively justified and are fairly easy to model, the bottom-line difference they make to our cost-effectiveness estimates is insufficient to warrant the (moderate) increase in the complexity of our models. These adjustments would also introduce an inconsistency between our methodologies for top charities. As a result, we are not planning to incorporate subnational adjustments at this time.

When would we revisit this conclusion?

We will revisit using subnational malaria mortality estimates if AMF or Malaria Consortium start working in countries where it would make a large difference to the bottom line. We would include subnational adjustments if AMF contributed nets in any of these countries: Djibouti (+500% adjustment), South Africa (+259%), and Swaziland (+126%), where malaria is endemic in some parts of the country but not others. We would also consider subnational adjustments if AMF contributed nets in Namibia (+25%), Kenya (+23%), Madagascar (+14%), or Rwanda (+10%).11The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_11").tooltip({ tip: "#footnote_plugin_tooltip_text_11", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

We will investigate whether subnational adjustments would make a substantial difference if Malaria Consortium enters additional countries; at this time, we do not have details on which regions are eligible for SMC in countries in which Malaria Consortium is not currently operating.12We have not yet prioritized getting details on which regions are eligible for SMC in countries in which Malaria Consortium does not currently work, as this would likely impose a substantial time cost on Malaria Consortium. jQuery("#footnote_plugin_tooltip_12").tooltip({ tip: "#footnote_plugin_tooltip_text_12", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

You can read the internal emails discussing our decision process here.

Notes   [ + ]

1. ↑ In both cases, we rely on reports by Cochrane, an organization that produces systematic reviews and other synthesized research to inform decision-makers. For AMF, we use a decline in all-cause mortality, because the Cochrane review of anti-malarial bed net distributions reports the effect in terms of a reduction in all-cause mortality. For Malaria Consortium, we use a decline in malaria mortality (proxied by a decline in malaria incidence), as the Cochrane review of seasonal malaria chemoprevention reports the effect in terms of a reduction in malaria incidence, but not all-cause mortality. See our cost-effectiveness analysis for more details. 2. ↑ We assume that the regional distribution of malaria incidence is a reasonable proxy for the regional distribution of malaria mortality. 3. ↑ A limitation of this analysis is it does not account for the possibility that AMF and Malaria Consortium are causing locations that are higher priority or lower priority than the average location already covered by nets or eligible for SMC to be covered on the margin. We do not explicitly include estimates of the marginal region funded in our cost-effectiveness analysis because we often have limited information about which regions would be covered with marginal additional funds. 4. ↑ We assume that where nets have been delivered in the past is a good proxy for where new nets will be delivered in the future. The data and calculations are in this spreadsheet. 5. ↑ See Cell J114. We did not include Papua New Guinea (where AMF funds some nets) in this analysis, as MAP only covers countries in Africa. 6. ↑ “The suitability of an area for SMC is determined by the seasonal pattern of rainfall, malaria transmission and the burden of malaria. SMC is recommended for deployment in areas: (i) where more than 60% of the annual incidence of malaria occurs within 4 months (ii) where there are measures of disease burden consistent with a high burden of malaria in children (incidence ≥ 10 cases of malaria among every 100 children during the transmission season) (iii) where SP and AQ [the drugs used to treat children] retain their antimalarial efficacy.” WHO SMC field guide (2013), Pg 8. 7, 11. ↑ The data and calculations are in this spreadsheet. 8. ↑ See Cell C126. 9. ↑ We use these scores as a qualitative guide to help us think through what to include in our cost-effectiveness analysis. You can see the rubric we use to assign scores in this spreadsheet. 10. ↑ You can read more about MAP’s methodology in this paper. 12. ↑ We have not yet prioritized getting details on which regions are eligible for SMC in countries in which Malaria Consortium does not currently work, as this would likely impose a substantial time cost on Malaria Consortium. function footnote_expand_reference_container() { jQuery("#footnote_references_container").show(); jQuery("#footnote_reference_container_collapse_button").text("-"); } function footnote_collapse_reference_container() { jQuery("#footnote_references_container").hide(); jQuery("#footnote_reference_container_collapse_button").text("+"); } function footnote_expand_collapse_reference_container() { if (jQuery("#footnote_references_container").is(":hidden")) { footnote_expand_reference_container(); } else { footnote_collapse_reference_container(); } } function footnote_moveToAnchor(p_str_TargetID) { footnote_expand_reference_container(); var l_obj_Target = jQuery("#" + p_str_TargetID); if(l_obj_Target.length) { jQuery('html, body').animate({ scrollTop: l_obj_Target.offset().top - window.innerHeight/2 }, 1000); } }

The post Why we don’t use subnational malaria mortality estimates in our cost-effectiveness models appeared first on The GiveWell Blog.

Summary

We recently completed a small project to determine whether using subnational baseline malaria mortality estimates would make a difference to our estimates of the cost-effectiveness of two of our top charities, the Against Malaria Foundation and Malaria Consortium. We ultimately decided not to include these adjustments because they added complexity to our models and would require frequent updating, while only making a small difference (a 3-4% improvement) to our bottom line.

Though this post is on a fairly narrow topic, we believe this example illustrates the principles we use to make decisions about what to include in our cost-effectiveness model.

Background

Two of our top charities—the Against Malaria Foundation (AMF) and Malaria Consortium’s seasonal malaria chemoprevention program—implement programs to prevent malaria, a leading killer of people in low- and middle-income countries.

One of the core reasons we recommend AMF and Malaria Consortium is their cost-effectiveness: how much impact they have (e.g., cases of malaria prevented, malaria deaths averted) with the funds they receive. Our estimates of charities’ cost-effectiveness isn’t just helpful to us in determining which charities should be GiveWell top charities; we also rely on these estimates to guide our decisions about how to allocate funding between our top charities.

Our cost-effectiveness estimates for AMF and Malaria Consortium use country-wide data on malaria mortality and malaria incidence in the places that both organizations work.1In both cases, we rely on reports by Cochrane, an organization that produces systematic reviews and other synthesized research to inform decision-makers. For AMF, we use a decline in all-cause mortality, because the Cochrane review of anti-malarial bed net distributions reports the effect in terms of a reduction in all-cause mortality. For Malaria Consortium, we use a decline in malaria mortality (proxied by a decline in malaria incidence), as the Cochrane review of seasonal malaria chemoprevention reports the effect in terms of a reduction in malaria incidence, but not all-cause mortality. See our cost-effectiveness analysis for more details. jQuery("#footnote_plugin_tooltip_1").tooltip({ tip: "#footnote_plugin_tooltip_text_1", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); However, neither organization serves a whole country—rather, they operate in sub-national regions—so the use of country-level estimates could cause us to either underestimate or overestimate their cost-effectiveness. If, for example, these programs are focused in the areas of the country with the highest malaria burden, using the average burden for the country would lead us to underestimate their cost-effectiveness. So, we completed a project to determine how much of an impact using subnational estimates would have, to consider whether we ought to incorporate this information into our cost-effectiveness analysis.

How we estimated the impact of subnational malaria incidence

AMF distributes insecticide-treated nets to prevent malaria; Malaria Consortium’s seasonal malaria chemoprevention (SMC) program provides preventive anti-malarial drugs. We used estimates of subnational malaria incidence from the Malaria Atlas Project (MAP) to see if regions covered by nets or eligible for SMC had higher or lower incidence than the average in the country in which they are located.2We assume that the regional distribution of malaria incidence is a reasonable proxy for the regional distribution of malaria mortality. jQuery("#footnote_plugin_tooltip_2").tooltip({ tip: "#footnote_plugin_tooltip_text_2", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

We focused on all areas covered by nets or eligible for SMC (rather than those covered by our top charities, specifically) for two reasons:

1. Our understanding is that when our top charities contribute resources to a country’s net distribution or SMC programs, the marginal region covered by these additional resources is not necessarily the same as the region to which these resources are assigned (because these resources are fungible with other resources within the national programs).3A limitation of this analysis is it does not account for the possibility that AMF and Malaria Consortium are causing locations that are higher priority or lower priority than the average location already covered by nets or eligible for SMC to be covered on the margin. We do not explicitly include estimates of the marginal region funded in our cost-effectiveness analysis because we often have limited information about which regions would be covered with marginal additional funds. jQuery("#footnote_plugin_tooltip_3").tooltip({ tip: "#footnote_plugin_tooltip_text_3", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });
2. Our aim is to estimate the cost-effectiveness of funds donated to these organizations in the future. The subnational region where AMF has worked in the past has not historically been a good indicator of the region where it will work in future.

Results for net distributions in countries where AMF works

We looked at geographical variation in malaria incidence in countries where AMF works, weighting each region by the number of nets it currently receives.4We assume that where nets have been delivered in the past is a good proxy for where new nets will be delivered in the future. The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_4").tooltip({ tip: "#footnote_plugin_tooltip_text_4", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

The average net delivered in the countries in which AMF works is hung in an area with 0-9% higher malaria incidence than the average in that country, and the weighted average adjustment to AMF’s cost-effectiveness would be 3% (in other words, AMF becomes 3% more cost-effective if we incorporate subnational estimates).5See Cell J114. We did not include Papua New Guinea (where AMF funds some nets) in this analysis, as MAP only covers countries in Africa. jQuery("#footnote_plugin_tooltip_5").tooltip({ tip: "#footnote_plugin_tooltip_text_5", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

Country Adjustment Zambia +9% Uganda +4% Ghana +4% Democratic Republic of the Congo +1% Togo +1% Malawi +0% Results for SMC in countries where Malaria Consortium works

We looked at six countries comprising >95% of Malaria Consortium’s SMC spending and compared malaria incidence in districts eligible for SMC with the country-wide average.6“The suitability of an area for SMC is determined by the seasonal pattern of rainfall, malaria transmission and the burden of malaria. SMC is recommended for deployment in areas: (i) where more than 60% of the annual incidence of malaria occurs within 4 months (ii) where there are measures of disease burden consistent with a high burden of malaria in children (incidence ≥ 10 cases of malaria among every 100 children during the transmission season) (iii) where SP and AQ [the drugs used to treat children] retain their antimalarial efficacy.” WHO SMC field guide (2013), Pg 8. jQuery("#footnote_plugin_tooltip_6").tooltip({ tip: "#footnote_plugin_tooltip_text_6", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });7The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_7").tooltip({ tip: "#footnote_plugin_tooltip_text_7", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

The average region eligible for SMC in countries where Malaria Consortium works has -2% to 17% higher malaria incidence than the average in that country. The weighted average adjustment to Malaria Consortium’s cost-effectiveness would be 4%.8See Cell C126. jQuery("#footnote_plugin_tooltip_8").tooltip({ tip: "#footnote_plugin_tooltip_text_8", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

Country Adjustment Commentary Guinea +17% Conakry, the capital, is ineligible for SMC and has low incidence. Nigeria +12% SMC appears to be targeted in the north, where malaria incidence is slightly higher. Niger +2% The majority of the population is either covered or planned to be covered from 2019. Burkina Faso 0% All districts are eligible. Mali 0% All districts are eligible. Chad -2% The four regions with very low malaria incidence (Borkou, Tibesti, Ennedi Est and Ouest) aren’t eligible for SMC, but are sparsely populated. What we concluded

We decided not to include these adjustments in our cost-effectiveness analysis because they increased complexity, without substantially affecting the bottom line.

When we decide whether to include adjustments in our model in general, we use a framework that first takes our best guess of the likely effect size and then rates each of the remaining question on a three-point scale.

Score9We use these scores as a qualitative guide to help us think through what to include in our cost-effectiveness analysis. You can see the rubric we use to assign scores in this spreadsheet. jQuery("#footnote_plugin_tooltip_9").tooltip({ tip: "#footnote_plugin_tooltip_text_9", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); Commentary Best guess of effect size 3-4% Can it be objectively justified? 3/3 While we have not investigated the MAP data in detail, we would guess that after further investigation, we would conclude it provides a reasonable approximation of subnational malaria incidence.10You can read more about MAP’s methodology in this paper. jQuery("#footnote_plugin_tooltip_10").tooltip({ tip: "#footnote_plugin_tooltip_text_10", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] }); How easily can it be modelled? 3/3 The methodology is clear and simple. Is it consistent with our other cost-effectiveness analyses? 2/3 We could include subnational adjustments for both of our top charities that implement malaria-prevention programs, but we believe it is unlikely there would be sufficient data to do the same for prevalence of worms or vitamin A deficiency (the focus of five of our other seven top charities).

Even though these adjustments can be objectively justified and are fairly easy to model, the bottom-line difference they make to our cost-effectiveness estimates is insufficient to warrant the (moderate) increase in the complexity of our models. These adjustments would also introduce an inconsistency between our methodologies for top charities. As a result, we are not planning to incorporate subnational adjustments at this time.

When would we revisit this conclusion?

We will revisit using subnational malaria mortality estimates if AMF or Malaria Consortium start working in countries where it would make a large difference to the bottom line. We would include subnational adjustments if AMF contributed nets in any of these countries: Djibouti (+500% adjustment), South Africa (+259%), and Swaziland (+126%), where malaria is endemic in some parts of the country but not others. We would also consider subnational adjustments if AMF contributed nets in Namibia (+25%), Kenya (+23%), Madagascar (+14%), or Rwanda (+10%).11The data and calculations are in this spreadsheet. jQuery("#footnote_plugin_tooltip_11").tooltip({ tip: "#footnote_plugin_tooltip_text_11", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

We will investigate whether subnational adjustments would make a substantial difference if Malaria Consortium enters additional countries; at this time, we do not have details on which regions are eligible for SMC in countries in which Malaria Consortium is not currently operating.12We have not yet prioritized getting details on which regions are eligible for SMC in countries in which Malaria Consortium does not currently work, as this would likely impose a substantial time cost on Malaria Consortium. jQuery("#footnote_plugin_tooltip_12").tooltip({ tip: "#footnote_plugin_tooltip_text_12", tipClass: "footnote_tooltip", effect: "fade", fadeOutSpeed: 100, predelay: 400, position: "top right", relative: true, offset: [10, 10] });

You can read the internal emails discussing our decision process here.

Notes   [ + ]

1. ↑ In both cases, we rely on reports by Cochrane, an organization that produces systematic reviews and other synthesized research to inform decision-makers. For AMF, we use a decline in all-cause mortality, because the Cochrane review of anti-malarial bed net distributions reports the effect in terms of a reduction in all-cause mortality. For Malaria Consortium, we use a decline in malaria mortality (proxied by a decline in malaria incidence), as the Cochrane review of seasonal malaria chemoprevention reports the effect in terms of a reduction in malaria incidence, but not all-cause mortality. See our cost-effectiveness analysis for more details. 2. ↑ We assume that the regional distribution of malaria incidence is a reasonable proxy for the regional distribution of malaria mortality. 3. ↑ A limitation of this analysis is it does not account for the possibility that AMF and Malaria Consortium are causing locations that are higher priority or lower priority than the average location already covered by nets or eligible for SMC to be covered on the margin. We do not explicitly include estimates of the marginal region funded in our cost-effectiveness analysis because we often have limited information about which regions would be covered with marginal additional funds. 4. ↑ We assume that where nets have been delivered in the past is a good proxy for where new nets will be delivered in the future. The data and calculations are in this spreadsheet. 5. ↑ See Cell J114. We did not include Papua New Guinea (where AMF funds some nets) in this analysis, as MAP only covers countries in Africa. 6. ↑ “The suitability of an area for SMC is determined by the seasonal pattern of rainfall, malaria transmission and the burden of malaria. SMC is recommended for deployment in areas: (i) where more than 60% of the annual incidence of malaria occurs within 4 months (ii) where there are measures of disease burden consistent with a high burden of malaria in children (incidence ≥ 10 cases of malaria among every 100 children during the transmission season) (iii) where SP and AQ [the drugs used to treat children] retain their antimalarial efficacy.” WHO SMC field guide (2013), Pg 8. 7, 11. ↑ The data and calculations are in this spreadsheet. 8. ↑ See Cell C126. 9. ↑ We use these scores as a qualitative guide to help us think through what to include in our cost-effectiveness analysis. You can see the rubric we use to assign scores in this spreadsheet. 10. ↑ You can read more about MAP’s methodology in this paper. 12. ↑ We have not yet prioritized getting details on which regions are eligible for SMC in countries in which Malaria Consortium does not currently work, as this would likely impose a substantial time cost on Malaria Consortium. function footnote_expand_reference_container() { jQuery("#footnote_references_container").show(); jQuery("#footnote_reference_container_collapse_button").text("-"); } function footnote_collapse_reference_container() { jQuery("#footnote_references_container").hide(); jQuery("#footnote_reference_container_collapse_button").text("+"); } function footnote_expand_collapse_reference_container() { if (jQuery("#footnote_references_container").is(":hidden")) { footnote_expand_reference_container(); } else { footnote_collapse_reference_container(); } } function footnote_moveToAnchor(p_str_TargetID) { footnote_expand_reference_container(); var l_obj_Target = jQuery("#" + p_str_TargetID); if(l_obj_Target.length) { jQuery('html, body').animate({ scrollTop: l_obj_Target.offset().top - window.innerHeight/2 }, 1000); } }

The post Why we don’t use subnational malaria mortality estimates in our cost-effectiveness models appeared first on The GiveWell Blog.

GiveWell is dedicated to finding outstanding giving opportunities and publishing the full details of our analysis. In addition to evaluations of other charities, we publish substantial evaluation of our own work. This post lays out highlights from our 2017 metrics report, which reviews what we know about how our research impacted donors. Please note:

• We report on “metrics years” that run from February through January; for example, our 2017 data cover February 1, 2017 through January 31, 2018.
• We differentiate between our traditional charity recommendations and the work of the Open Philanthropy Project, which became a separate organization in 2017 and whose work we exclude from this report.
• More context on the relationships between GiveWell, Good Ventures, and the Open Philanthropy Project can be found here.

Summary of influence: In 2017, GiveWell influenced charitable giving in several ways. The following table summarizes our understanding of this influence.

Headline money moved: In 2017, we tracked $117.5 million in money moved to our recommended charities. Our money moved only includes donations that we are confident were influenced by our recommendations.

Money moved by charity: Our nine top charities received the majority of our money moved. Our seven standout charities received a total of $1.8 million.

Money moved by size of donor: In 2017, the number of donors and amount donated increased across each donor size category, with the notable exception of donations from donors giving $1,000,000 or more. In 2017, 90% of our money moved (excluding Good Ventures) came from 20% of our donors, who gave $1,000 or more.

Donor retention: The total number of donors who gave to our recommended charities or to GiveWell unrestricted increased about 29% year-over-year to 23,049 in 2017. This included 14,653 donors who gave for the first time. Among all donors who gave in the previous year, about 42% gave again in 2017, up from about 35% who gave again in 2016.

Our retention was stronger among donors who gave larger amounts or who first gave to our recommendations prior to 2015. Of larger donors (those who gave $10,000 or more in either of the last two years), about 73% who gave in 2016 gave again in 2017.

GiveWell’s expenses: GiveWell’s total operating expenses in 2017 were $4.6 million. Our expenses decreased from about $5.5 million in 2016 due to the Open Philanthropy Project becoming a separate organization in June 2017. We estimate that 67% of our total expenses ($3.1 million) supported our traditional top charity work and about 33% supported the Open Philanthropy Project. In 2016, we estimated that expenses for our traditional top charity work were about $2.0 million.

Donations supporting GiveWell’s operations: GiveWell raised $5.7 million in unrestricted funding (which we use to support our operations) in 2017, compared to $5.6 million in 2016. Our major institutional supporters and the six largest individual donors contributed about 49% of GiveWell’s operational funding in 2017.

Web traffic: The number of unique visitors to our website remained flat in 2017 compared to 2016 (when excluding visitors driven by AdWords, Google’s online advertising product).

For more detail, see our full metrics report (PDF).

The post GiveWell’s money moved and web traffic in 2017 appeared first on The GiveWell Blog.